# EkLine > EkLine finds what AI agents get wrong about your product and fixes what they read. It measures what answer engines say about you, traces each answer back to the pages behind it, and opens pull requests that correct them across docs, help center, and marketing. Buyers ask Claude, ChatGPT, Gemini, or Perplexity what to use, then ask Cursor, Claude Code, Codex, or GitHub Copilot to build with it. EkLine works on both halves of that. It checks what your team and your coding agents write before it merges, and it drafts what nobody has written yet from the changes you ship. Every change arrives as a pull request your team approves; EkLine never publishes on its own. ## Docs Agent Reads your merged changes, works out which pages they touched, and drafts the updates as a pull request. Also generates documentation from a codebase, a video, or sources you already have. - [Docs Agent](https://ekline.io/solutions/docs-automation): the product page: how the agent researches, verifies and ships documentation updates, with customer results - [Overview](https://docs.ekline.io/agent/): what the Docs Agent does and how a run works - [Getting started](https://docs.ekline.io/agent/getting-started/): connect a repository and open a first pull request - [Create documentation](https://docs.ekline.io/agent/create/): generate guides and reference pages from a codebase or a video - [Generate an API reference](https://docs.ekline.io/agent/generate-api-reference/): build reference pages from an OpenAPI spec - [Write release notes](https://docs.ekline.io/agent/release-notes/): turn merged pull requests into a release note - [Refresh documentation screenshots](https://docs.ekline.io/agent/refresh-doc-screenshots/): recapture screenshots when the UI changes - [Prevent documentation drift](https://docs.ekline.io/agent/prevent-documentation-drift/): catch pages that a shipped change made wrong - [Manage the knowledge base](https://docs.ekline.io/agent/manage-knowledge-base/): the sources the agent reads before it writes - [Agent integrations](https://docs.ekline.io/agent/integrations/): the tools the Docs Agent connects to ## Docs Reviewer Checks documentation in CI before it merges, including the documentation coding agents write. Comments on the pull request. - [Overview](https://docs.ekline.io/reviewer/overview/): what gets checked and what a finding looks like - [Quickstart](https://docs.ekline.io/reviewer/quickstart/): every way to run the reviewer - [GitHub Action](https://docs.ekline.io/reviewer/quickstart/github-action/): run the reviewer on pull requests - [CLI](https://docs.ekline.io/reviewer/quickstart/cli/): run the reviewer locally or in any CI system - [VS Code extension](https://docs.ekline.io/reviewer/quickstart/vscode/): findings inline as you write - [Google Docs add-on](https://docs.ekline.io/reviewer/quickstart/google-docs/): the same checks on Google Docs - [Style guides](https://docs.ekline.io/reviewer/configuration/style-guides/): bring your own style guide and terminology - [Automatic pull request review](https://docs.ekline.io/agent/automatic-pr-review/): review every documentation change without asking ## Integrations - [All integrations](https://ekline.io/integrations): source control and CI (GitHub, GitLab); coding agents and editors (Claude, Cursor, VS Code); documentation frameworks (Astro, Docusaurus, Fern, Gatsby JS, GitBook, Mintlify, ReadMe, Redocly); content and support (Confluence, Google Drive, Notion, Pylon, Slack, Zendesk); issue tracking and analytics (Jira, Linear, PostHog) - [GitHub](https://docs.ekline.io/reviewer/integrations/github-integration/): review and enforce documentation standards in pull requests - [GitLab](https://docs.ekline.io/reviewer/integrations/gitlab-integration/): the same checks on merge requests - [Bitbucket](https://docs.ekline.io/reviewer/integrations/bitbucket-integration/): the same checks on Bitbucket pull requests ## Security and privacy - [Trust center](https://ekline.io/trust-center): SOC 2 Type 2 compliance, and the subprocessors EkLine uses with where each one is located ## Frequently asked questions ### What is EkLine? EkLine is a set of AI agents for developer tool and API companies. They check what ChatGPT and Claude say about your product and whether coding agents like Cursor can build with it, then fix the docs, help center, and marketing pages those agents read, using your code, pull requests, and support tickets. Every change arrives as a pull request or draft your team approves. Teams at Socure, P0 Security, and ContextQA use EkLine. ### How is EkLine different from SEO, GEO, and AI visibility tools like Profound or Semrush? Visibility tools report how often AI mentions you and suggest what to change. EkLine makes the change: it reads your code, pull requests, docs, and support tickets, writes the missing or wrong pages, and ships them into your stack for your team to approve. It is built for developer tools and APIs, so it also checks whether coding agents can build with your product. Keep your SEO tool; EkLine works alongside it. ### Why not just ask Claude or Cursor to do this? You can, for one page, once. EkLine does it on every release. It reads the merged pull requests, tickets, and Slack threads, finds every page the change touched across docs, help center, and blog, drafts in your style guide, and opens the pull request. It also checks what Claude Code and Cursor write in your CI, and shows whether agents now answer correctly. ### Do we have to move off Mintlify, ReadMe, or GitBook? No. EkLine works with the stack you already have: Mintlify, ReadMe, GitBook, Docusaurus, Confluence, Notion, Zendesk, Pylon, or Markdown in your repo. Your docs platform’s AI only works on the docs it hosts. EkLine works across every surface agents read, from docs to help center to blog, and your content stays in your stack. If your docs live in Git, setup takes about 15 minutes. ### How do we know what EkLine writes is accurate? EkLine writes from your own sources: your code, merged pull requests, tickets, and existing docs. Every draft shows the sources it used and is checked against your style guide and terminology. ContextQA’s engineers merge 97% of what EkLine writes. ### What can EkLine access, and what happens to our data? EkLine only accesses the repositories and tools you connect, with the permissions you grant. It opens pull requests and drafts, and nothing merges or publishes without your approval. Each customer is isolated, agent runs are sandboxed with scoped credentials, and your data is never used to train models. EkLine is SOC 2 compliant. ### How do we measure whether EkLine is working? EkLine starts with a baseline: what agents say about you, which pages they cite, and whether coding agents finish the tasks your developers care about. Re-run the same questions and tasks after you ship and compare, then count what merged. Socure saw nearly 10x more documentation PRs than its previous monthly average. ## Company - [Website](https://ekline.io): what EkLine does, in full - [Pricing](https://ekline.io/pricing): plans and what each one includes - [Customers](https://ekline.io/case-study): case studies with numbers - [ContextQA](https://ekline.io/case-study/how-contextqas-engineers-merge-97-percent-of-what-an-ai-agent-writes): how ContextQA's engineers came to merge 97.5% of what an AI agent wrote - [P0 Security](https://ekline.io/case-study/how-p0-securitys-engineering-team-stopped-writing-docs-and-started-shipping-faster): how P0 Security's engineering team stopped writing docs and started shipping faster - [Socure](https://ekline.io/case-study/how-socure-built-developer-experience-into-a-moat-with-ekline): how Socure built developer experience into a moat with EkLine - [Documentation](https://docs.ekline.io): the full product documentation site - [Blog](https://ekline.io/blog): writing on documentation, GEO, and AI agents ## Optional - [Documentation index for LLMs](https://docs.ekline.io/llms.txt): the docs site's own llms.txt, generated from the same source as the pages - [Full documentation for LLMs](https://docs.ekline.io/llms-full.txt): every documentation page inlined — large, skip when context is tight - [Abridged documentation for LLMs](https://docs.ekline.io/llms-small.txt): the same, with non-essential content removed