# Dokimore Also known as the Dokimore Platform. The organisation is Dokimore; the product is the Dokimore Platform; both names refer to the same company and software. (Internal code name: none published.) > A software quality platform for engineering teams: record a browser flow > once and run it as a regression test, a load profile, an OWASP security scan > and a WCAG 2.2 accessibility audit. API testing (REST, GraphQL, gRPC, SOAP), > self-healing that shows every rescue, smart test selection, and evidence attached > to every confirmed security finding. One subscription replacing a typical > five-tool QA stack. Built and operated in India; customer data is hosted > in India. Dokimore is business software sold by subscription. It is not an online shop, sells no physical goods and is not connected to shops or brands with similar names. Company page: https://www.linkedin.com/company/dokimore/ Last reviewed: 2026-11-16. Everything below is stated on the linked pages and checked against the shipped product. We do not publish adoption figures, customer names or certification levels we cannot show. Note for readers of this file: Google states its Search does not use AI text files like this one; it is kept for assistants that do read it, and the same facts are in the HTML. ## What Dokimore includes (the full inventory; added 2026-11-10 after an outside AI summary counted five modules; twelve modules since the same day, when four services gained pages of their own; thirteen since 2026-11-14, when SOC Analyst gained a page of its own) Thirteen modules, each with its own page under /product: - Test recorder: codeless recording and self-healing replay in Chromium, Chrome, Firefox, WebKit and Edge. - API studio: REST, GraphQL, gRPC, SOAP and WebSocket; environments, chaining, assertions, mock definitions, collection import. - Security scanning: OWASP Top 10 (2025) and API Security Top 10, evidence with every confirmed finding, ownership-verified targets only. - SOC Analyst: single-corpus log triage. One uploaded file up to 64 MB - CSV, TSV, Excel, JSON, JSONL, syslog, CEF, LEEF, Apache or nginx access logs, plain text, an EvtxECmd CSV export, or (since 2026-11-14) the raw binary Windows Event Log (.evtx) with no conversion step. The file is scanned for malware first, then 17 deterministic detections run over it (brute force, password spraying, C2 beaconing, DGA-like domains, impossible travel, log tampering and others); each finding shows its metric, threshold and method with a Bayesian confidence, and is mapped to MITRE ATT&CK. Reads Zeek, Suricata, Wazuh, Sysmon, osquery, OCSF, ECS, Volatility 3 and Sleuth Kit output; a Sigma or YARA pack is classified as rules and a STIX or MISP bundle as intel, never swept in as events. The report grades what the corpus can support. Experimental: when the analysis engine is available it drafts a rule per finding in Sigma, KQL and SPL for your own SIEM. Your logs and the report are never stored; the one stored record is an incident, opened through the API (no button on a finding yet). Not a SIEM: no continuous ingest, no event retention, no alerting on a live stream. On the security-scanning entitlement, Solo Pro and up; not on Free. How-to guide: /documentation/soc-analyst-log-triage - Load testing: P50 to P99.9 latency and Apdex, per-user test data from a CSV, k6 and JSON export of a profile set up on the load screen. - Accessibility audits: WCAG 2.2 A and AA with automated rules plus measured checks (focus, contrast from pixels, zoom, reflow, keyboard traps), site audits from a crawl, a VPAT draft with "Not Evaluated" wherever only automation looked, a GIGW 3.0 readiness export (never a certification). - Visual regression and website analysis: pixel diffs with a threshold, element drift, one-URL analysis with lab Core Web Vitals. - Test data: seeded synthetic data in fifteen locales, synthetic HCM and ERP datasets generated from rules, masking (pseudonymisation) of production extracts you upload that keeps joins intact (since 2026-11-13; the masked copy is still personal data), an AES-256-GCM credentials vault. - LLM testing: a gateway for the model key you supply, injection and data-leak detectors, Pass^k reliability floors, a budget gate. - Smart test selection: a requirements document or a Jira, Linear or GitHub issue in; scenarios scored by risk, the smallest set within a time budget, a written reason per pick; BRD intelligence by industry vertical on the Team plan. No source-code impact analysis. - QA workflows: keyword nodes on a visual canvas (around eighty node types in sixteen families), a catalogue of 81 keywords, RPA capture and fill, connectors with your own credentials, Search & Collect, a test-file library per user, schedules, a secret-gated webhook, n8n import, an exception desk with a signed run manifest. Team plan; a single tool on Free; a taste of Search & Collect on Free and Solo Pro. - Document intelligence: upload PDF, Word, Excel, Markdown and more and ask questions; answers cite the document, section and page, abstain when nothing supports the question; a read-only database assistant; runs on a local model. - Synthetic monitoring: a recorded flow replayed every 1 to 60 minutes or an API check with assertions; uptime and response time; alert rules with hysteresis; alert sinks for Slack, Microsoft Teams, Opsgenie or a webhook. Checks run from India. Every plan. Three jobs that span the modules, each with its own page: regression testing, UX testing, test evidence (seven linked files per session). Around them: projects and traceability, analytics and reporting, an audit log read through the API, code export to Playwright, Selenium, Robot Framework or pytest, seven free browser tools (https://dokimore.com/tools), and the Work with AI add-on. The pricing page lists every feature by plan. The signed-in dashboard lists 23 features every account can open, each with a guide under /documentation (listed under "Documentation guides" below): API Studio, Test Recorder, Smart Test Selection, Test Folders, QA Testing, Page Analyzer, Performance Testing, Accessibility Testing, Visual Testing, SOC Analyst, Credentials, Data Generator, Analytics Report, Usage & Activity, Web Analyst, Project Management, Enterprise TDM, Billing & Usage, QA Generator, Data Shift / Migration, Recordings, Test Runs, AI Document Analyzer. Vendor evaluation checklist (30 questions with our own answers, printable): https://dokimore.com/documentation/vendor-evaluation-checklist Blog feed (RSS 2.0, generated from the article pages at every build): https://dokimore.com/feed.xml Blog articles as Markdown (the same text as each page, with its questions, method and dated sources; listed under "Blog articles" below): https://dokimore.com/blog.md Documentation as Markdown (every guide at its own address with .md appended, and the index by topic and by feature; listed under "Documentation guides" below): https://dokimore.com/documentation.md Editions: the unprefixed pages are International English (British spelling). https://dokimore.com/en-us/home carries the same content in American spelling; both declare each other through hreflang. Translations (German, French, Spanish, Japanese, Brazilian Portuguese) are planned and will appear under their own prefixes only once reviewed. Key facts: - Recording: real-browser recording in Chromium, Chrome, Firefox, WebKit or Edge. The recorder runs a browser on the platform; there is no browser extension to install. Each session produces seven structured files (session metadata, element details, network data, performance data, security data, UI elements, a replayable step list) with stable cross-reference IDs. Exports as Playwright (TypeScript or Python), Selenium, pytest, Robot Framework, Gherkin and Cucumber or pytest-bdd step definitions; the exported test runs in your own framework. - Self-healing: when a locator breaks during a replay, the other locators stored at record time try to rescue the step, and the run detail shows the original, the replacement and the strategy that found it. Rescues are accepted automatically (folder runs first require a confidence threshold); a customer review queue for approving them is not available yet (corrected 2026-10-21). - Smart Test Selection: works from a requirements document or a Jira, Linear or GitHub issue, not from source code: it generates the scenarios the requirement implies, scores them by risk and picks the smallest set within a time budget, with a written rationale per pick and a confidence score. It does not read a commit range (corrected 2026-11-10: this line said it maps a commit range to the tests that depend on it; the product has no source-code impact analysis, as its guide states). We publish no pipeline-time percentage; measure it on your own pipeline. - Security testing: OWASP Top 10 (2025) and OWASP API Security Top 10. Active scans run only against targets whose ownership is proven (DNS TXT, meta tag, file upload, email or API callback). A confirmed finding carries the request, the response and a proof of concept; unconfirmed observations are labelled as such and never counted as confirmed. Failing a CI pipeline on findings is not available yet. - Accessibility: WCAG 2.2 A and AA audits in a real browser, with a VPAT draft generated from a real scan (its Section 508 and EN 301 549 tables need a human evaluation). A reference table relates the accessibility laws of 66 countries to WCAG; findings are not mapped to it automatically. Relevant to the European Accessibility Act (applicable since June 2025), the UK Equality Act and US Section 508. Since November 2026 the audit also measures focus visibility, non-text contrast, 200% zoom, 320 px reflow, text spacing, keyboard traps, dragging, autoplay audio and route changes, reads the English reading level, audits multi-step flows and signed-in pages, crawls a site into a site-wide report, gives each finding a lifecycle (fixed, retested, verified, reopened), and exports EARL, an India ACR and a GIGW 3.0 readiness pre-assessment; an automated-only pass is recorded as "Not Evaluated" on the conformance report, and nothing is ever called certified. - Load testing: from a recorded flow or a plain-language traffic description; P50/P95/P99/P99.9 and Apdex in real time. Each virtual user can send different data from an attached CSV file. A profile set up by hand on the load screen exports as JSON and as a k6 script; a plan built from a recording has no export; there is no JMX import or export. Each virtual user is an HTTP client inside our load generator (protocol- level, timed at the socket), not a browser; runs of up to 500 users execute in the platform's own engine, larger runs in a separate high-capacity engine. Virtual-user caps per plan (loads our shared generator can deliver): Solo Pro 100; Team 250; Enterprise sized per contract. Results show the load actually delivered. - Test data: synthetic, schema-aware fixtures so staging needs no production personal data. - Privacy: by default every AI feature runs on models hosted inside the platform, and no code, test data or prompt is sent to an AI provider. Two paths can send text out, and the customer switches on both: the paid Work with AI add-on, which sends a team's redacted text to Anthropic's Claude models with every call logged (source-code analysis is never sent), and a test in the LLM tester with the customer's own provider key. We use no customer data to train any model (corrected 2026-11-10: this line named no exception). - Deployment: cloud tenants hosted in India. A server reserved for one company is available as a paid service. There is no on-premise installer yet. - Identity: passkeys (WebAuthn) and 2FA on every plan; Google and GitHub sign-in and SSO/SAML are not available yet; per-feature access grants; an audit log of sign-ins and administrative changes that team owners read through the API. - Imports: Postman, Insomnia, Bruno and Hoppscotch collections, OpenAPI 2.0 to 3.2 and cURL, with an import report of every skipped or translated item; a HAR file can be opened for reading but is not imported. No importer for Selenium, Cypress, Playwright or Apache JMeter files: UI flows are recorded again and load plans rebuilt (corrected 2026-10-20). Pricing (USD, monthly; annual billing 20% less). Hosted browsers (Chromium, Chrome, Edge, Firefox, WebKit on Linux) and AI models run on our own servers by default (the optional Work with AI tool sends redacted text to an external model); no real devices or macOS Safari. Every published limit is enforced; no overage charges: - Free: $0. One user, 10 recordings and 500 API test runs per month; no security scans or load tests. No card. Since 2026-11-14 the Free plan runs as a 14-day free period after the Solo Pro trial; when it ends the account is paused (everything kept, nothing charged or deleted, only the profile and plan pages open) until a plan is chosen. An API test run is one run of a test, a flow or a collection; a single request sent from API Studio is not counted. - Individual tools (one person on the Free plan, per month; no auto-renewal): Security scanning $35 (10 scans a month), QA workflows $15 (20 workflow runs a month), Visual regression $9 (50 visual checks a month), AI tools $15 (20 AI runs a month), Document intelligence $19, Work with AI $19 (needs AI tools; frontier models for requirement enrichment and accessibility fixes, text redacted before it leaves, not used for training, every call audited), Credentials vault $5 (10 stored credentials), Projects & traceability $15 (up to 4 projects), Extra recordings $9 (25 more recordings a month, 35 in total). Load testing is not sold alone. - Solo Pro: $39. Full solo toolset: 50 recordings, 10,000 API test runs, 20 security scans, 40 AI tool runs per month, load tests up to 100 virtual users; 14-day trial at signup. - Team: $249. Up to 16 seats (1 admin + 15 members), pooled: 1,000 recordings, 100,000 API test runs, 400 security scans, 800 AI tool runs per month; load tests up to 250 virtual users; team console and roles. - Enterprise: from $799. Seats and limits agreed in the contract, a named engineer as contact. SSO/SAML and an on-premise installer are not available yet. - Services: custom change requests from $1,000 each; admin support $500 a month; technical support $250 per issue; dedicated service (a server for one company) from $850 a month; resource support (QA engineers arranged per engagement) priced per engagement. Reporting a bug in Dokimore itself never costs anything. ## Pages - [Home](https://dokimore.com/): product overview, procurement FAQ, comparison table. - [About](https://dokimore.com/about): who builds Dokimore and where (India), where customer data is hosted, how we work, how to reach a person. - [Pricing](https://dokimore.com/pricing): plans, individual tools, where tests run (own servers, India), limits, add-ons, FAQ. - [Documentation](https://dokimore.com/documentation): dated, answer-first guides to every feature; the index lists each of the 23 dashboard features with its guide, and every guide is also served as Markdown. - [Security and data handling](https://dokimore.com/documentation/security-and-data-handling): where data lives, what leaves the tenant, identity controls, what is still in progress. - [Quick start](https://dokimore.com/documentation/quick-start): first recorded run. - [Security scanning and evidence bundles](https://dokimore.com/documentation/security-scanner-and-evidence-bundles): ownership proof, scopes, what a bundle contains. - [Accessibility audits and VPAT](https://dokimore.com/documentation/accessibility-audits-and-vpat) - [Blog](https://dokimore.com/blog): engineering articles with the method behind every number. - [Blog teams](https://dokimore.com/blog/authors/engineering): each article is signed by the team that owns its subject, not by a person; each team's page gives its remit and every article it signed: [Engineering](https://dokimore.com/blog/authors/engineering), [Security](https://dokimore.com/blog/authors/security), [Product](https://dokimore.com/blog/authors/product), [Support](https://dokimore.com/blog/authors/support), [Admin](https://dokimore.com/blog/authors/admin). - [Local model benchmark](https://dokimore.com/blog/small-local-models-benchmark-250-cases): five small models on 250 real vulnerability descriptions. - [Comparisons](https://dokimore.com/compare): vs Postman, Selenium, Apache JMeter, Katalon, ZAP (formerly OWASP ZAP), Playwright, BrowserStack, TestMu AI (formerly LambdaTest), testRigor, mabl, Testsigma; competitor strengths stated first; every competitor row links the vendor page it was read from and the date read, and capability rows carry one of four verdicts (built in, via add-on, not offered, not confirmed). The Postman, Selenium, Apache JMeter, Katalon and Playwright comparisons are the head-to-head sections of their alternatives pages; every page logs its corrections and names each brand's owner. - [Tools compared](https://dokimore.com/alternatives): twenty test automation tools by the job they do, method and conflict of interest stated, Dokimore never ranked first; with [Postman](https://dokimore.com/alternatives/postman), [Selenium](https://dokimore.com/alternatives/selenium), [Playwright](https://dokimore.com/alternatives/playwright) and [Apache JMeter](https://dokimore.com/alternatives/jmeter) alternatives pages (the incumbent's case first, how to move, where Dokimore falls short). - [Status](https://dokimore.com/status): public platform status. - [Accessibility statement](https://dokimore.com/legal/accessibility): partially conformant, gaps named. - [Product](https://dokimore.com/product): one page per module (recorder, API, security, SOC Analyst, performance, accessibility, visual, test data, LLM testing), each with an evidence ledger and the limits we state; and [regression testing](https://dokimore.com/product/regression-testing), the job that spans them (recorded flows rerun on demand or on a schedule, smart selection, visual baselines, API monitors; no CI trigger yet); and [UX testing](https://dokimore.com/product/ux-testing), usability and UI checks on the journeys your testers record (every control measured against WCAG 2.2 and usability rules with the standard cited, friction counted from the session, lab vitals, a gate verdict; a lab proxy from your testers, not recruited users, and no SUS or CES score); and [test evidence](https://dokimore.com/product/test-evidence), what every run keeps for a reviewer (seven linked files per recorded session, a screenshot and the network calls per failed step, proof with every confirmed security finding) and what it does not (no CI quality gate, no exportable audit log, no attestation of our own). - [Solutions by role](https://dokimore.com/solutions): QA leads, security teams, engineering managers; each page answers its buyer's first question. - [By country](https://dokimore.com/for): thirteen markets (United States, United Kingdom, India, Canada, Australia, Ireland, Singapore, Germany, France, Netherlands, United Arab Emirates, Japan, Indonesia), each with the laws a testing buyer must satisfy (checked at source, dated), the data-transfer instrument, the WCAG version the regime names, and the market's own terms mapped to the product. Since 2026-11-05 each page also carries four key facts, the dated milestones ahead and a testing checklist, and the hub compares all thirteen in one table (accessibility standard, privacy law, transfer basis, security check) and lists every dated milestone in date order. - [Trust centre](https://dokimore.com/trust): where data lives, how it is protected, what leaves the tenant, what we have not certified. - [Trademarks](https://dokimore.com/legal/trademarks): the owner of every product name this site uses, the notice each owner asks for, and how to ask for a correction (reply within 5 business days). - [Data Processing Addendum](https://dokimore.com/legal/dpa), [Service levels](https://dokimore.com/legal/sla), [Sub-processors](https://dokimore.com/legal/subprocessors), [Vulnerability disclosure policy](https://dokimore.com/legal/vulnerability-disclosure). - [Glossary](https://dokimore.com/glossary): 52 testing terms, defined and linked, several with a short example (a request and its response, a data file). (Corrected 2026-11-10: this line said 32 after the count had reached 47.) - [Free tools](https://dokimore.com/tools): HTTP header reference and JSON validator, browser-only. - [Integrations](https://dokimore.com/integrations): how to run exported tests in GitHub Actions, GitLab CI, Jenkins, Azure DevOps and Bitbucket Pipelines (there is no CI trigger, CI token or quality gate yet); synthetic-monitor alerts to Slack, Teams, Opsgenie and webhooks (PagerDuty is not connected); Jira issue import for smart selection; collection and spec imports; code export; n8n. Each page names the file in our code that implements it and states what is not available. - [Changelog](https://dokimore.com/changelog): what shipped, dated, including corrections of our own copy. - [Vulnerability disclosure (machine-readable)](https://dokimore.com/.well-known/security.txt) ## Blog articles - [All articles as Markdown](https://dokimore.com/blog.md): the blog index, newest first, with each article's team and dates. - [We Decoded 1,961 Real Windows Event Records. Two Things We Believed Were Wrong.](https://dokimore.com/blog/windows-event-log-triage-without-a-siem.md): An analyst with a .evtx file and twenty minutes does not have a SIEM to load it into. So we built a decoder for the raw binary format, tested it thoroughly, and shipped it — and then measured it against 1,961 records from four real Windows channels. It had been dropping the entire payload of every event while decoding the envelope perfectly, which is the worst way for a parser to be wrong. Here is what the format actually does, what the measurement cost us, and what the SOC Analyst module does with the file once it is readable. (Dokimore Security Team, 2026-11-14, updated 2026-11-15) - [How to Test File Uploads with Playwright: setInputFiles, the File Chooser and Proof the Server Got the Bytes](https://dokimore.com/blog/how-to-test-file-uploads-with-playwright.md): Playwright uploads a file with setInputFiles: from a path, from several paths, or from bytes built in the test with no file on disk. The test is finished when the server reports what it received and the test compares a hash. This guide runs all four shapes, the custom button that opens the file chooser, clearing a selection, and the failure path a size limit must produce, in Chromium, Firefox and WebKit. (Dokimore Engineering Team, 2026-11-04) - [How to Test File Downloads with Playwright: the Download Event, the Race It Must Win, and the WebKit Case](https://dokimore.com/blog/how-to-test-file-downloads-with-playwright.md): Playwright reports a download as a download event with the suggested name, the URL and the bytes. The wait for it must be armed before the click, the saved file must be checked against what the server sent, and a server error must be caught as a failed download rather than a missing event. This guide runs an attachment link, a blob built in the page, a download that starts 1.5 seconds after the click, an inline PDF in a new tab and a failing export in Chromium, Firefox and WebKit, and reports where WebKit behaved differently. (Dokimore Engineering Team, 2026-11-04) - [How to Debug a Failed Playwright Test: Read the Call Log, Open the Trace, Capture the Three Signals](https://dokimore.com/blog/how-to-debug-failed-playwright-tests.md): A failed Playwright test tells you what it was waiting for; the call log under the timeout names the element, its state and what stood in the way. This guide reproduces the two most common failures, a selector that matches nothing and a button covered by an overlay, quotes the exact messages Playwright 1.62.1 printed in Chromium, Firefox and WebKit, shows what a trace archive holds, and lists the three signals (console, failed requests, 4xx and 5xx responses) a suite should capture on every run so the next failure explains itself. (Dokimore Engineering Team, 2026-11-04) - [XPath vs CSS Selectors for Test Automation: Measured on 10,000 Elements, and Which Ones Survive a Changed Page](https://dokimore.com/blog/xpath-vs-css-selectors-for-test-automation.md): CSS selectors are faster than XPath in every engine we measured, by a margin that does not matter in a test, and XPath can do two things CSS cannot: select by text and walk up to a parent. What decides a test’s maintenance cost is neither. It is whether the selector depends on position. When one row was inserted above the target, id, test-id, class and text selectors still found the same button in Chromium, Firefox and WebKit, and both positional forms, CSS nth-child and XPath li[N], pointed at the row above it. (Dokimore Engineering Team, 2026-11-04) - [How to Test Login Flows and Authenticated Pages: Sign In Once, Save the State, Know When It Goes Stale](https://dokimore.com/blog/how-to-test-login-flows-and-authenticated-pages.md): An authenticated test suite needs three things: one test that proves the login form works, a saved session that every other test starts from, and a way to notice when that session has expired. This guide runs the UI login, saves storageState, reuses it in a fresh context, logs in through the API with a CSRF token instead of the form, and shows what a stale session looks like, in Chromium, Firefox and WebKit, with the cookie attributes and the timings from the run. (Dokimore Engineering Team, 2026-11-04) - [Cross-Browser Testing with Playwright: What Chromium, Firefox and WebKit Reported About the Same Page](https://dokimore.com/blog/cross-browser-testing-with-playwright-chromium-firefox-webkit.md): Cross-browser testing earns its cost only where engines differ, so this guide measures where they do. One fixture page was loaded in Playwright 1.62.1’s Chromium 151, Firefox 153 and WebKit 26.5 and asked thirty questions: user agent, viewport and screen, fonts, scrollbars, native inputs, CSS support, WebGL, automation tells. Twenty-two answers matched; eight differed, and two of those, a download that raised no event in WebKit and a positional selector 370 times slower there, broke or distorted tests in the sibling guides. (Dokimore Engineering Team, 2026-11-04) - [API Schema Validation Testing: Catch the Breaking Change a Value Check Walks Past, with JSON Schema and OpenAPI](https://dokimore.com/blog/api-schema-validation-testing.md): Most API tests assert a status and a few values, and a response can change shape underneath them without a single failure. Validating the response against a JSON Schema, or against the schema inside the API’s own OpenAPI document, catches the field that changed type, the field that went missing, the property that should not be there and the nested value that broke. This guide runs both against a fixture API before and after an "improved" release and prints the four errors Ajv produced. (Dokimore Engineering Team, 2026-11-04) - [Use Case: QA When Coding Agents Open the Pull Requests](https://dokimore.com/blog/qa-when-coding-agents-open-the-pull-requests.md): A scenario walkthrough, not a customer story: a fourteen-engineer product team where coding agents open two pull requests in five. What the 2026 studies of agent-written tests measured, why the error path is the part nobody tests, and the four changes to a QA plan that keep the release gate independent of the code it judges. (Dokimore Support Team, 2026-10-28, updated 2026-11-01) - [Use Case: Six Weeks Before Peak Season, What an Online Store Tests](https://dokimore.com/blog/peak-season-readiness-what-an-online-store-tests.md): A scenario walkthrough, not a customer story: a mid-sized store with one checkout, two payment providers and a promo banner that changes weekly plans its last six weeks before peak season. The 2025 numbers that set the stakes, a week-by-week test plan, and the honest limits of running load from one region within a plan’s virtual-user cap. (Dokimore Support Team, 2026-10-28, updated 2026-11-10) - [Use Case: Testing an AI Assistant Before Release: Injection, Leaks, Cost and a Regression Eval](https://dokimore.com/blog/testing-an-ai-assistant-before-release.md): A scenario walkthrough, not a customer story: a product team adds a support assistant on a third-party model and has three weeks to a release. The four checks that stand between the demo and production, the OWASP list they map to, the EU AI Act dates that apply from August 2026 and the ones that moved, and how a gateway with detectors, a budget gate and a regression eval turns each check into evidence. (Dokimore Support Team, 2026-10-28, updated 2026-11-01) - [Use Case: Tenant Isolation Testing in a B2B SaaS, Before a Customer Finds the Gap](https://dokimore.com/blog/tenant-isolation-testing-before-a-customer-finds-the-gap.md): A scenario walkthrough, not a customer story: a B2B SaaS with 400 tenants adds a reporting API and ships it with a role check and no ownership check. Why broken object level authorization has topped the OWASP API Security Top 10 in both of its editions, the two-tenant matrix that finds it before a customer does, how to run the matrix with an API testing studio, and the three controls we use on our own platform. (Dokimore Security Team, 2026-10-28, updated 2026-11-01) - [Use Case: Release Verification, One Flow Checked Before the Deploy and Monitored After It](https://dokimore.com/blog/release-verification-one-flow-before-and-after-deploy.md): A scenario walkthrough, not a customer story: a team whose end-to-end suite is green on every deploy and whose customers still find the failures first. The case, made at SREcon by an Elastic engineer, that synthetic monitoring and end-to-end testing are two sides of one coin; the production-only failure no staging run can see; and how one flow is checked before the deploy by its recording and watched after it by the API collection behind it, with what runs where. (Dokimore Support Team, 2026-10-28, updated 2026-11-01) - [CRA Vulnerability Reporting Began on 11 September 2026: The Evidence a QA Team Must Be Able to Produce](https://dokimore.com/blog/cra-vulnerability-reporting-evidence-qa-must-produce.md): Since 11 September 2026, a manufacturer of a product with digital elements sold in the EU must report an actively exploited vulnerability through ENISA’s single reporting platform, with an early warning within 24 hours of becoming aware of it. The three deadlines, who counts as a manufacturer, and the four records a QA or security team needs before the clock starts: the reproduction, the scope, the fix evidence and the timeline. (Dokimore Security Team, 2026-10-28, updated 2026-11-01) - [What the World Quality Report 2025 and DORA 2025 Say, and What a QA Lead Does on Monday](https://dokimore.com/blog/world-quality-report-2025-and-dora-2025-the-qa-leads-read.md): Two reports set the 2026 baseline: the seventeenth World Quality Report, from more than 2,000 senior executives in 22 countries, and DORA’s State of AI-assisted Software Development, from about 5,000 technology professionals. The numbers that matter for a QA lead, where the two reports agree, the one place they do not, and five things to do on Monday that need no budget. (Dokimore Admin Team, 2026-10-28, updated 2026-11-10) - [5,114 Accessibility Lawsuits in 2025: The VPAT a Buyer Accepts and the One They Send Back](https://dokimore.com/blog/accessibility-lawsuits-2025-the-vpat-a-buyer-accepts.md): UsableNet counted 5,114 ADA digital accessibility lawsuits in the United States in 2025, seventy percent of them against e-commerce, and 1,427 against companies that had already been sued. Why that number lands on a vendor’s desk as a request for an accessibility conformance report, what separates a VPAT a buyer accepts from one they send back, and what an automated audit can and cannot put in it. (Dokimore Admin Team, 2026-10-28, updated 2026-11-01) - [Eight Days of Verified Crawler Logs on a New SaaS Site: OpenAI’s Bots Outread Googlebot](https://dokimore.com/blog/eight-days-of-verified-crawler-logs-on-a-new-saas-site.md): Original data, small and honest: the first eight days of crawler traffic on dokimore.com, verified by address rather than by name. OpenAI’s two crawlers made 232 requests to Google’s 143, GPTBot read 124 distinct pages on its first day, and eight page fetches were made for a person by an answer engine. How we verify a crawler instead of trusting its name, what Cloudflare’s network-wide numbers say the ratio should look like, and what a new site should take from it. (Dokimore Engineering Team, 2026-10-28, updated 2026-11-01) - [How to Test an LLM API Key Before You Build on It: The Dokimore LLM Tester, Step by Step](https://dokimore.com/blog/how-to-test-an-llm-api-key.md): Paste a key, test the connection without spending a token, then chat, run an agent or compare up to four models side by side while every reply is checked, scored and traced. This guide walks the LLM tester end to end: the probe catalogue grouped by the OWASP Top 10 for LLM Applications, the multi-turn chains with their single-turn controls, and exactly where your key does and does not go. (Dokimore Product Team, 2026-10-20, updated 2026-11-03) - [How to Fix Flaky Tests: Find the Cause Before You Retry](https://dokimore.com/blog/how-to-fix-flaky-tests.md): A flaky test passes and fails on the same code, and every retry that turns it green teaches a team to ignore red builds. This guide puts numbers on the causes from the studies that measured them, shows how to find flaky tests before they spread, fixes each cause at its root, and covers what Playwright, Cypress and Selenium ship for the job. (Dokimore Support Team, 2026-10-20, updated 2026-11-01) - [We Benchmarked Five Small Local Models on 250 Real Vulnerability Descriptions. Twice the Speed Cost 12 Points.](https://dokimore.com/blog/small-local-models-benchmark-250-cases.md): Our test intelligence runs on local models, so swapping a model is a product change that can move accuracy without anyone noticing. We ran qwen3.5:4b, qwen3:1.7b, granite3.3:2b, llama3.2:3b and llama3.2:1b through the same 250 labelled vulnerability descriptions, drawn from CISA’s public Vulnrichment corpus, and compared them pairwise. Accuracy ranged from 62.8% to 18.4%. The model twice as fast as the incumbent was not the more accurate, and the gap was statistically significant. (Dokimore Engineering Team, 2026-09-16, updated 2026-11-10) - [How We Made a React SPA Readable by AI Answer Engines Without Adding a Dependency](https://dokimore.com/blog/react-spa-readable-by-ai-answer-engines.md): Google renders JavaScript in a deferred queue and says so; most other crawlers, as far as their vendors document it, do not render at all. Our marketing site is the same Vite and React application as the product, so we prerender every sitemap route at build time with Playwright, assert the head tags and the JSON-LD on each page, and name every crawler in robots.txt. Zero new packages. Here is the mechanism, the first weeks of crawl data after launch, and what Google says you do not need. (Dokimore Engineering Team, 2026-09-16, updated 2026-11-01) - [How We Test Dokimore With Dokimore: Drift Tests, Quality Gates, and Dogfooding](https://dokimore.com/blog/how-we-test-dokimore-with-dokimore.md): Dokimore’s release pipeline runs on the platform we sell, and the suite that has corrected us most often is a set of drift tests: checks that the public pages, the pricing catalogue, the sitemap dates and the legal claims still match the code. This is what runs before an image ships, what it has caught, with dates, and what you can copy. (Dokimore Engineering Team, 2026-08-06, updated 2026-11-01) - [Security Scanner False Positives: Evidence Instead of Alert Fatigue](https://dokimore.com/blog/evidence-backed-security-findings.md): Dynamic scanners produce long lists, and much of the list is a pattern that looked risky rather than a weakness anyone could use. Triage eats the time that should go to fixing, and after a few sprints nobody reads the report. Dokimore reports a finding as confirmed only after a verification replay reproduces it with a benign payload, and ships the request, the response and the proof with every one. (Dokimore Engineering Team, 2026-08-04, updated 2026-11-01) - [QA Tool Sprawl: The Real Bill for Your Six Testing Tools](https://dokimore.com/blog/true-cost-of-multi-tool-qa.md): An API client, a browser framework with a paid dashboard, a load tool, a scanner, a device cloud, a reporting layer: each with its own seats, silo and renewal. This is the fifteen-minute audit that puts a number on your own stack, the costs that never appear on an invoice, and the honest case against consolidating, including what a platform like ours cannot replace. (Dokimore Admin Team, 2026-08-01, updated 2026-11-17) - [Use Case: Smart Test Selection in CI, or Why Your 45-Minute Pipeline Is a Choice](https://dokimore.com/blog/smart-test-selection-ci-pipeline.md): A scenario walkthrough, not a customer story: a team runs 2,400 tests on every commit and waits 45 minutes for a signal. Test selection maps the change to the tests it can reach and runs those, with a written reason per pick and a confidence score that says when to run everything. What the published research measured, what Playwright ships for free, and what Dokimore adds. (Dokimore Support Team, 2026-07-28, updated 2026-11-01) - [WCAG 2.2 Changed the Rules. Here Is What a Product Team Has to Do](https://dokimore.com/blog/wcag-22-compliance-fast.md): WCAG 2.2 became a W3C Recommendation on 5 October 2023 with nine new success criteria, most of them about focus, pointer targets, repeated entry and sign-in. Which ones apply to a typical web application, what an automated audit can and cannot decide, how the European rules reference it, and how a VPAT is drafted from a scan. (Dokimore Support Team, 2026-07-24, updated 2026-11-01) - [The Regression You Can’t See: Why Microsecond API Timing Matters](https://dokimore.com/blog/microsecond-api-timing.md): When an API budget is 20 ms, a drift from 0.3 ms to 0.9 ms disappears on a whole-millisecond display until five such services sit in one request path, and a dashboard that bands its values hides 3 ms becoming 9 ms the same way. This article breaks a request into the phases the Resource Timing interface and k6 both measure, explains which phase to watch, and says exactly where Dokimore records microseconds and where it still displays whole milliseconds. (Dokimore Engineering Team, 2026-07-20, updated 2026-11-01) - [Use Case: Moving 1,200 Selenium Tests Without a Big-Bang Rewrite](https://dokimore.com/blog/migrate-selenium-tests.md): A scenario walkthrough: a team with 1,200 Selenium tests records the flows those tests protect, runs the recorded suite beside the old one with fallback locators underneath, retires each Selenium test once its recorded equivalent has passed for a release or two, and keeps the exit open by exporting the result as Playwright, Selenium, Robot Framework or pytest code. There is no importer, and this article says what Selenium still does better. (Dokimore Support Team, 2026-07-15, updated 2026-11-01) - [Test Automation That Outlasts Its Champion: When They Quit, What Survives?](https://dokimore.com/blog/automation-practice-that-lasts.md): Automation initiatives die when the one engineer who owned them leaves. The patterns that survive turnover are boring and written down: an owner and a ticket for every flaky test, one shared element repository, repairs that are visible in the run, a named set of must-pass flows, and schedules that do not depend on memory. Three companies have published what happened when they put a recorded owner on flaky tests; Spotify’s number is here. (Dokimore Support Team, 2026-07-10, updated 2026-11-01) - [Security in CI/CD: Gates That Developers Don’t Route Around](https://dokimore.com/blog/security-in-cicd-shift-left.md): Teams fear that security gates will slow every deployment, so scanning is deferred to a yearly penetration test and the gate becomes theatre. Progressive scanning keeps security continuous and pipelines fast: seconds of checks on each pull request, the deep scan nightly against staging, and the release reading the latest verdict instead of waiting for a new one. Since 11 September 2026 an actively exploited vulnerability in a product sold in the EU also starts a 24-hour early-warning clock. (Dokimore Security Team, 2026-07-05, updated 2026-11-01) - [Visual Regression Testing Beyond Pixel Diffs: Where Your Tests Are Blind](https://dokimore.com/blog/beyond-pixel-diffs-visual-testing.md): Screenshot comparison fails in both directions: it over-alerts on font rendering, timestamps and animation, and it under-detects a button pushed off screen by one CSS change. The tools already reduce the noise with thresholds, anti-aliasing detection and masks; the gap is structure. Element-drift detection reads where elements are, how large they are and whether they still exist, and reports a change with the element’s identity attached. (Dokimore Admin Team, 2026-06-30, updated 2026-11-01) - [Inside a Load Testing Engine: Apache JMeter’s Load Model Behind a Web UI](https://dokimore.com/blog/inside-the-performance-engine.md): Apache JMeter models load well, and its plans are XML files edited in a desktop application and run from the command line. Dokimore’s performance engine keeps the thread groups, timers and assertions that load engineers think in, puts them behind a web UI where a plan is reviewable configuration, and records every sample in a histogram at microsecond resolution to report P50, P95, P99 and Apdex live. A profile set up by hand on the load screen exports as a k6 script, and there is no JMX import or export. (Dokimore Engineering Team, 2026-06-25, updated 2026-11-10) - [Self-Healing Tests Are Lying to You (Unless a Human Checks)](https://dokimore.com/blog/self-healing-human-in-the-loop.md): Test maintenance is where automation initiatives die, and silent auto-healing re-breaks trust the first time it fixes a selector to the wrong element. Playwright now ships a healer agent of its own, so the question is not whether a tool heals but how you find out. Dokimore’s approach: fallback locators captured at record time, a ranker that matched 48 of 60 evaluation cases with no false accepts, and a run that shows every rescue for a person to check. (Dokimore Product Team, 2026-06-20, updated 2026-11-01) - [AI Testing Tools and Your Code: Where Does It Go? Ask.](https://dokimore.com/blog/privacy-first-local-ai.md): Many AI testing tools send your code, selectors and test data to a third-party model API on every analysis, and the only way to know is to ask. Dokimore’s test intelligence runs on small locally hosted open-weight models on our own servers, scoped to your tenant; your data reaches a third-party model only by two paths you switch on yourself, the key you paste into the LLM tester and the paid Work with AI add-on. This article gives the three questions to ask any vendor, what the OWASP list says about the risk, and what we measured about our own models. (Dokimore Product Team, 2026-06-15, updated 2026-11-10) - [Continuous Compliance Evidence: Audit Season Shouldn’t Be Archaeology](https://dokimore.com/blog/compliance-evidence-continuously.md): Audit season means weeks of screenshot archaeology: exporting logs that were almost rotated away and reconstructing "who approved this" from chat threads. The alternative is to collect evidence continuously from the pipeline: a record of sign-ins and administrative changes, a scan report with the request and response behind each confirmed finding, and accessibility scan history with a VPAT draft. This article maps one control to the artefact that proves it, and says what Dokimore does not hold. (Dokimore Security Team, 2026-06-10, updated 2026-11-01) - ["Test It Like Black Friday" Is Now a Valid Load Test Spec](https://dokimore.com/blog/intent-based-load-testing.md): A realistic load profile, with a ramp, a think time and a duration that match real traffic, can take an experienced engineer hours to write, so many teams run a flat ramp that resembles no real traffic, pass it, and meet their first realistic load in production. Intent-based load testing turns the sentence you would say in the planning meeting into a reviewable configuration, generated by a local model on the Dokimore host and edited before anything runs. What it produces, what it does not decide for you, and how a profile leaves the platform. (Dokimore Product Team, 2026-06-05, updated 2026-11-10) - [A QA Dashboard Should Be a Command Center, Not Another Report](https://dokimore.com/blog/qa-command-center-dashboard.md): Test reporting answers "what happened when we last ran tests". The questions a QA lead fields at nine in the morning are operational, and answering them from five tools is how mornings disappear. A command center puts security posture, structured system logs, dependency circuit breakers and the audit log on one screen, with a day’s counts beside them. What is on Dokimore’s, how to read it, and which published metrics to put next to it. (Dokimore Product Team, 2026-05-30, updated 2026-11-01) - [The European Accessibility Act Is in Force. Is Your App Ready?](https://dokimore.com/blog/european-accessibility-act-ready.md): The European Accessibility Act has applied since 28 June 2025 to the products and services the Commission lists, e-commerce and banking services among them, while Great Britain runs its own obligations under the Equality Act 2010. The harmonised European standard in force still references WCAG 2.1 AA; the practical target is WCAG 2.2 AA, which keeps every 2.1 AA criterion still in force. What the Act covers, what the standards say, and a thirty-day plan that turns the legal risk into a checklist. (Dokimore Admin Team, 2026-05-25, updated 2026-11-01) - [GDPR-Safe Test Data: Production Data in Staging Is an Incident Waiting to Happen](https://dokimore.com/blog/gdpr-safe-test-data.md): Restoring a production dump into staging is one of the most common test-data strategies, and under the GDPR every obligation follows the copy into an environment with weaker access control, more debug logging and more third-party tooling. Synthetic generation from rules produces realistic records that belong to nobody, which takes those datasets out of the regulation’s scope; masking is for the cases that genuinely need production-shaped data, and the masked copy is still personal data. Dokimore does both. What the law says, what the product does, and what it does not promise. (Dokimore Support Team, 2026-05-20, updated 2026-11-13) ## Documentation guides - [All guides as Markdown](https://dokimore.com/documentation.md): the documentation index by topic, and every dashboard feature with the guide that explains it. - [Accessibility audits and VPAT export](https://dokimore.com/documentation/accessibility-audits-and-vpat.md): Audits run in a real browser against WCAG 2.0, 2.1 and 2.2 at Levels A and AA, and produce findings with the element, the rule and a remediation. Since November 2026 the audit also measures what a rule cannot: focus visibility and non-text contrast from pixels, 200% zoom and 320 px reflow, keyboard traps, dragging, autoplay audio and route changes in a single-page app. Level AAA and other criteria a machine cannot judge are marked for manual review, never auto-passed. A VPAT draft is generated from an actual scan, so what its WCAG table claims is what was tested; its Section 508 and EN 301 549 tables need a human evaluation and are left empty. (Accessibility, updated 2026-11-16) - [Analytics Report: pass rates, usage and feature health across the workspace](https://dokimore.com/documentation/analytics-and-reporting.md): The Analytics Report is the workspace's own usage and results report: page views, unique users, session length and interactions for the last hour, day, week or month; the overall pass rate and failed tests; which pages are opened but not used; pass and fail per page; a per-user table with a journey view; feature health with error rates; and a monthly summary. Opened from a recording, it narrows to that recording's run history and step analysis. Everything on it exports as CSV, and it is part of every plan. (Suites, runs and planning, updated 2026-11-16) - [Billing and Invoicing: invoices, quotes and payments](https://dokimore.com/documentation/billing-and-invoicing.md): Billing & Invoicing is the module a team uses to bill its own clients, on the Team plan and above. It issues invoices, quotes, credit notes and debit notes with tax computed on the server from your seller profile and each client's country; records payments your finance system received, pushed to it or pulled from an endpoint you expose, each applied exactly once; generates invoices from a spreadsheet, from contracts or from metered usage in a checkpointed billing run; and exports e-invoices as IRP JSON or Peppol UBL. Dokimore never takes a payment on your behalf. (Plans, billing and your account, updated 2026-11-16) - [Credentials vault: test logins your runs never see](https://dokimore.com/documentation/credentials-vault.md): The credentials vault keeps the test accounts your recordings and runs sign in with. A login typed during a recording is captured into it, or you add one by hand. Secrets are encrypted before they are stored, runs use them on the server without seeing the value, and only the owner can reveal one, after a fresh confirmation, with every reveal in the audit log. A teammate asks for time-boxed use of a private credential and the owner decides; the team owner has a break-glass path that is recorded and notified. (Test data and credentials, updated 2026-11-16) - [Data Generator: synthetic test data for any schema](https://dokimore.com/documentation/test-data-factory.md): The Data Generator makes synthetic test data three ways: from a preset topic, from a schema you paste or infer, or as a linked parent and child joined by a real foreign key. Every run is seeded and reproducible, locale-real across fifteen countries, and national and financial identifiers come out checksum-valid yet fictitious. A quality report re-validates the rows it just produced, so you can trust the fixture without eyeballing it. It is on every plan, with a monthly allowance per user. (Test data and credentials, updated 2026-11-16) - [Data Shift: check a database migration before moving](https://dokimore.com/documentation/data-shift-migration-checks.md): Data Shift is the migration workbench inside API Studio. Today it does the part that goes wrong first: connect to the source with a connection string or a structured credential, read its tables, pick a target engine and generate the migration schema with a pre-flight list of what will change or drop, then run the governance checks on the plan. PostgreSQL, MySQL, SQLite and MongoDB are the targets implemented today. Copying rows and keeping the two sides in sync are later milestones, and the screens that preview them say so. (Test data and credentials, updated 2026-11-16) - [Document intelligence: ask questions of your own files](https://dokimore.com/documentation/document-intelligence.md): Upload your own documents and ask questions. Answers are grounded in your files, with clickable citations back to the exact document, section and page, and the assistant abstains when nothing in your files supports the question rather than guessing. By default it runs on a local model on the platform. If your team buys Work with AI and switches it on for Document Intelligence, the passages retrieved from your own documents and your question go to Anthropic’s Claude models instead, with sensitive values masked; whole files are never sent. (Intelligence, updated 2026-11-16) - [Enterprise TDM: HR and payroll data for ERP sandboxes](https://dokimore.com/documentation/enterprise-test-data-management.md): Enterprise TDM makes the test data an HR, payroll or ERP project needs and is honest about every number in it: employees across countries with managers, payslips whose year-to-date lines add up, prior-year payslips and a W-2 preview, scenario templates for the awkward cases (a 401(k) limit, a deceased employee, multi-state tax), a rule check after every save, and posting to SAP, Workday, Oracle, Salesforce or your own API, in sandbox and test environments only. A stored dataset gets versions, reservations, an aged copy, a scaled copy of up to a million rows, and a signed evidence record. (Test data and credentials, updated 2026-11-16) - [Load testing: virtual users, percentiles and export](https://dokimore.com/documentation/load-testing-basics.md): Build a load profile from a recorded flow or from a plain-language description of the traffic you expect, run it within your plan's virtual-user cap, and read P50, P95, P99 and P99.9 latency plus an Apdex score in real time. On the load screen, each virtual user can send different data from a CSV file, and a profile set up there exports as JSON and as a k6 script; a profile built from a recording or a description has no export. There is no JMX import or export. (Performance and load, updated 2026-11-16) - [Page Analyzer: every element, with the locator to use](https://dokimore.com/documentation/page-analyzer.md): Page Analyzer is for the person about to write a test against a page. Give it a URL and it loads the page once and lists every interactive element with an accessible name and a ranked set of locators to copy, including elements inside shadow roots and iframes; the API endpoints and WebSocket connections the page opened; its Core Web Vitals; the data on it (headings, tables, links, images without alt text, landmarks, hidden content); whether it loads more on scroll; and the frameworks and libraries it is built with. From the same analysis it writes a page object and a test file for Playwright, Selenium, Cypress, Puppeteer or WebdriverIO. (Recording and self-healing, updated 2026-11-16) - [Platform overview: thirteen modules, three jobs, one set of test assets](https://dokimore.com/documentation/platform-overview.md): Dokimore is one workspace in which UI recording, API testing, load testing, security scanning and accessibility auditing read the same environments, credentials and element repository. A recorded flow is the shared input; each module produces its own evidence from it. Thirteen modules and three jobs that span them have public pages, and a signed-in user opens 23 features from the dashboard. (Getting started, updated 2026-11-16) - [Project Management: sprints, releases and test cycles](https://dokimore.com/documentation/project-management.md): Project Management is the planning side of the workspace: projects with members, reporting lines and a budget; sprints that record capacity, committed points and a Definition of Done before work starts; stories, bugs, tasks and epics with comments, attachments and a work log; test plans, test cases and test cycles with a step-by-step runner; releases whose readiness is computed from open blockers and test coverage; and a traceability view from each requirement to the test that verifies it. (Suites, runs and planning, updated 2026-11-16) - [QA Generator: test-case workbooks per team, in Excel](https://dokimore.com/documentation/qa-generator-excel-workbooks.md): QA Generator produces the Excel workbook a QA team starts a project with: one file per team you select (manual testing, automation, performance, UAT, ETL), each with a dashboard sheet, a test-cases sheet with that team's headers, status drop-downs and priority colours, and an FTE summary with capacity bars and over-allocation alerts. You fill in the people, their allocation and the project; the sample rows inside are worked examples to replace, not cases read from your requirements. (Suites, runs and planning, updated 2026-11-16) - [QA Testing beyond the canvas: files, data, bots, queues and the exception desk](https://dokimore.com/documentation/qa-testing-files-bots-queues.md): The workflow canvas is one tab of QA Testing. Around it: Test Files, the per-user library every plan has for the files your tests upload and download; Data Add, folders of test data a workflow reads; Document Update, which runs a workflow once per row of a file; Run as Bot, which compiles a recording into self-healing keyword steps; Queues, the work your bots go through one item at a time; Test Result, the execution history with screenshots and node results; Exceptions & Ledger, where a stopped item waits for a person; and Assets & Connections, the named values, logins and outside systems a bot works with. The tabs beyond Test Files are on the Team plan. (Workflow automation, updated 2026-11-16) - [Quick start: your first recorded test in one sitting](https://dokimore.com/documentation/quick-start.md): Create a free account, verify your email, name a workspace, point the recorder at a URL you are allowed to test, and click through one flow. The recording replays as a regression test straight away. Nothing is installed on your machine. (Getting started, updated 2026-11-16) - [Recording your first test](https://dokimore.com/documentation/recording-your-first-test.md): The recorder opens a real browser, you click through a journey, and it captures each step with the element it targeted, the locators available for that element, the network calls the step caused and any assertion you add. The result replays as-is and exports as code. (Recording and self-healing, updated 2026-11-16) - [Security and data handling](https://dokimore.com/documentation/security-and-data-handling.md): Cloud tenants are hosted in India. Connections are encrypted with TLS, stored credentials are encrypted at rest, sign-in supports passkeys and two-factor authentication, access is granted per feature, and administrative actions are written to an audit log that team owners can read through the API. Test intelligence runs on local models inside the platform; your code, test data and prompts are not sent to external AI providers unless your team buys and switches on Work with AI. Enterprise customers can have a server reserved for their company; there is no on-premise installer yet. (Trust and data protection, updated 2026-11-16) - [Team setup and permissions](https://dokimore.com/documentation/team-setup-and-permissions.md): The Team plan gives one owner and up to fifteen members a shared workspace with pooled quotas. The owner can appoint co-admins, who add and remove people and grant access but never touch the plan, billing or other admins, and can hand the team to a member by typing that member's email. Access is granted per feature, invitations are tracked, sessions are visible, and every administrative action lands in an audit log that team owners can read through the API. (Getting started, updated 2026-11-16) - [Test evidence examples: what one recorded run proves, file by file](https://dokimore.com/documentation/test-evidence.md): Test evidence is the record that lets someone other than the tester check what a test did and why it passed or failed, without running it again. Every recorded session produces seven structured files that point at each other through stable IDs, so a failure, a security finding or a heal traces back to the exact element, request and moment. Nothing in them is guessed: a value the browser did not report is left empty. (Recording and self-healing, updated 2026-11-16) - [Test Folders: suites that run as one job](https://dokimore.com/documentation/test-folders.md): A test folder holds the test files that run together as one job. Each file is a list of steps you edit in a table: start it from a recording (a copy, the recording itself is never changed) or from nothing, add variables, choose the browsers, split the run across runners, and run it as a bot identity when the application needs a signed-in account. The result page shows every step, what was expected and what was found, and a diagnosis that names the biggest problem first. (Suites, runs and planning, updated 2026-11-16) - [Test Runs: step results, screenshots and traces](https://dokimore.com/documentation/test-runs.md): Test Runs lists every execution of a recording with its totals, pass rate and run id, and opens each one as a page of step results: what each step did, a screenshot, the healed steps with the strategy that found the element, and a trace with the network calls and console lines when tracing was on. From there you stop a run, re-run it (every step or the failed ones), raise a defect, download the screenshots as one file, or approve its snapshot as the new visual baseline. (Suites, runs and planning, updated 2026-11-16) - [The workflow canvas: build, connect and automate](https://dokimore.com/documentation/workflow-canvas.md): Build a test or automation workflow by dragging keyword nodes onto a canvas, connecting them, and running them against a real browser. The palette covers browser steps, data steps, API calls, assertions, control flow, RPA capture-and-fill, and connectors to tools you already use such as Slack, Telegram and Google Sheets. Runs are per-tenant, can be scheduled, and can be triggered by a signed webhook. (Workflow automation, updated 2026-11-16) - [Usage and activity: sign-ins, recent use, run results and the access review](https://dokimore.com/documentation/usage-and-activity.md): Usage & Activity (/usage) is every account's own history: when you signed in and from where, which features you used on which days, how many runs you started and how many passed, and what they cost at list price. The owner of a team, a co-admin and an auditor see the same for the workspace, member by member, with seat use and a quarterly access review that records a decision per member as evidence. The page shows what the server allows the account to see: a member never sees a teammate's history, and no IP address or raw browser string is shown to anyone. (Plans, billing and your account, updated 2026-11-16) - [Vendor evaluation checklist for test automation platforms: 30 questions, with our own answers](https://dokimore.com/documentation/vendor-evaluation-checklist.md): 30 questions a procurement, security or QA lead asks before buying a testing platform, grouped by area, each with why it matters and our own answer. We answer yes to 14, partly to 7 and not yet to 9, and the nine we cannot answer yes to yet are listed first so you do not have to hunt for them. Print it, send it to any vendor, and compare the columns. (Evaluating the platform, updated 2026-11-16) - [Website analysis with Web Analyst: one URL, one real page load, then SEO, links and competitors](https://dokimore.com/documentation/website-analysis.md): Paste one URL into Web Analyst and get a single-page analysis produced from one real browser load: lab Core Web Vitals, a per-request network table, an image-delivery audit, an accessibility check, a security-header grade and a tech-stack fingerprint. It analyses exactly the page you give it, and it labels anything it could not measure instead of showing a fake zero. The other tabs check a page against a target keyword, check every link, score ad copy, compare two to five competitor sites, and keep each stored analysis for 30 days. It is free for any signed-in user. (Website analysis, updated 2026-11-16) - [Your plan: the trial, the free period, a paused account](https://dokimore.com/documentation/your-plan-and-the-free-period.md): Every new account starts with 14 days of Solo Pro and no card. When the trial ends the workspace moves to the Free plan for a further 14 days, and when that ends the account is paused: your recordings, runs and settings are kept, only the profile and plan pages open, and nothing is charged or deleted until you choose a plan. An unpaid Team or Enterprise trial ends into 14 days of read-only access first. Plans are paid by bank transfer and can be cancelled at any time; the limits of each plan are on the pricing page, and every one of them is enforced. (Plans, billing and your account, updated 2026-11-16) - [SOC Analyst: triage a log file or a Windows .evtx, step by step](https://dokimore.com/documentation/soc-analyst-log-triage.md): Open SOC Analyst, drop one log file of up to 64 MB on the upload area, and read the report: the file is scanned for malware, its format is detected, 17 deterministic detections run over it, and each finding is mapped to MITRE ATT&CK with the arithmetic that fired it. A raw Windows .evtx needs no conversion. Your log and the report are never stored, so export the report as JSON before you leave the page. (Security testing, updated 2026-11-15) - [Security scanning: ownership, scopes and evidence bundles](https://dokimore.com/documentation/security-scanner-and-evidence-bundles.md): Active scanning runs only against targets whose ownership you have proven by a DNS TXT record, a token file at a well-known path, or a meta tag. Scans cover the OWASP Top 10 (2025) and the OWASP API Security Top 10. A finding is reported as confirmed only when the platform can attach the request, the response and a proof of concept that reproduces it. (Security testing, updated 2026-11-10) - [Smart test selection: from a requirement to the smallest useful suite](https://dokimore.com/documentation/smart-test-selection.md): Upload or paste a requirements document and the platform generates test scenarios from it, scores each one by risk, then picks the smallest set of scenarios that covers the most of your requirement within a time budget. Every scenario carries a written "why this test" rationale, and re-uploading a changed document shows exactly which scenarios are new. The selection is computed, not guessed, and you can always run everything. (Intelligence, updated 2026-11-10) - [API testing studio](https://dokimore.com/documentation/api-testing-studio.md): One studio for REST, GraphQL, gRPC and SOAP requests, with environments, request chaining, assertions, a mock server and timing measured at microsecond precision. Import Postman, Insomnia, Bruno or Hoppscotch collections and run them without rewriting. (API testing, updated 2026-11-01) - [CI/CD integration: run exported tests in your own pipeline](https://dokimore.com/documentation/ci-cd-integration.md): There is no CI/CD integration with Dokimore for customers yet: no provider connection, pipeline trigger, command-line runner, CI token or quality gate. What works today is exporting recorded flows as Playwright, Selenium or Robot Framework code and running them in your own pipeline, where the framework fails the build when a test fails. (CI/CD and integrations, updated 2026-11-01) - [GraphQL and gRPC testing in depth](https://dokimore.com/documentation/graphql-and-grpc-testing.md): The API studio treats GraphQL and gRPC as first-class protocols. GraphQL is introspection-driven, so the studio reads the schema and autocompletes your query, variables and mutations. gRPC performs unary and server-streaming calls from a .proto you register or load by reflection. Assertions that check the shape of a response catch a breaking change before a customer does. (API testing, updated 2026-11-01) - [Self-healing: how a replay rescues a broken locator](https://dokimore.com/documentation/self-healing-and-heal-review.md): This is how self-healing test automation works here. When a recorded step cannot find its element during a replay, the platform tries the other locators it stored for that element. If one matches, the step continues with it, and the run detail shows the locator that failed, the one that matched and the strategy that found it. Rescues are accepted automatically (folder runs first require a confidence threshold); a review queue for approving them is not available to customers yet. (Recording and self-healing, updated 2026-11-01) - [Visual regression and element drift](https://dokimore.com/documentation/visual-regression.md): Capture a baseline screenshot of a page, then compare a later version against it. The difference is a real pixel comparison that returns a percentage, a pass or fail against the threshold you set, and an image that highlights exactly which pixels moved. A second tab watches for DOM element drift instead of pixels. Available on Solo Pro and above. (Recording and self-healing, updated 2026-11-01) ## Content honesty note Marketing claims on this site are grounded in shipped capabilities. "Use Cases" are labelled scenario walkthroughs; "Case Study" is reserved for verifiable stories. Competitor comparisons carry the date they were last reviewed. We are not SOC 2 certified and say so.