Git-backed docs built for humans and AI agents
DocsAlot is a hosted docs-as-code platform that publishes human-readable documentation and packages the same source for AI agents, generating llms.txt, skill.md, and hosted MCP endpoints from one maintained docs layer. Teams connect a repo or existing sources (GitHub docs, help centers, API references, Notion), review diffs, and publish without maintaining a docs framework or deployment pipeline. The platform also includes AI visibility audits and benchmark reports showing what AI tools can actually find and cite.
Connect your repo via CLI, a coding-agent prompt (works with Codex, Cursor, Claude, and other repo-aware agents), or a web editor, then publish polished docs without maintaining a framework or deployment pipeline.
Every project generates llms.txt, skill.md, clean Markdown with stable anchors, and MCP-ready chunks from the same maintained source used for the human docs site.
Each project gets a hosted MCP server (e.g., mcp.docsalot.dev/yourco) with search_docs for retrieval, get_page for exact source text, and run_example for guided examples — no infrastructure to wire or maintain.
Every publish creates a shareable benchmark report showing which answers cite the docs, which sources agents use, and where onboarding gaps remain, such as drifted API examples or stale support answers.
Pulls from GitHub docs (MDX), help-center articles (HTML), OpenAPI API references, and product notes (Notion), plus Intercom, Zendesk, Confluence, and Google Docs via the done-for-you Docs Pack.
Embeddable AI widget for help centers, included from the Startup plan, with AI suggestions inside the editor for drafting and upkeep.
Team plan adds private docs, authenticated help centers, custom domain and subfolder hosting, and preview/review/publish workflows with unlimited publish actions.
Bring in GitHub docs, help-center articles, API references (OpenAPI), READMEs, changelogs, and product notes from tools like Notion.
DocsAlot generates stable anchors, clean markdown, llms.txt, skill.md, and MCP-ready chunks from the same source.
Ship hosted docs, review diffs, and audit what AI tools can actually see before onboarding breaks.
Plans start at $79/month (Startup) for one public docs site with AI-ready docs and MCP, $149/month (Team) for production hosting on a custom domain with private docs and 250 AI messages/month, and custom Enterprise pricing for governance, SSO, migration, and audit services. Add-ons include a Docs Pack for migration support, an AI Visibility Pack, and developer marketing add-ons. No per-seat charge
$79/month
One public docs site with AI-ready docs and MCP on startup traffic capacity, hosted MCP endpoint, core AI drafting and upkeep, AI suggestions in the editor, basic source integrations, AI help-center widget, 2 included languages, no per-seat
$149/month
Everything in Startup plus production hosting, higher docs and MCP traffic capacity, custom domain and subfolder hosting, private docs and authenticated help centers, 250 included AI messages/month, unlimited publish actions, 5 included lan
Custom
Everything in Team plus unlimited seats, contact-only rollout planning, self-correcting docs and automated drift checks, detailed analytics with account-level visitor identification, lead-gen signals, SSO (SAML) and SCIM, dedicated onboardi
llms.txt is an emerging standard: a compact, structured table of contents at your site root that gives LLMs a canonical map of your docs. When someone asks ChatGPT, Claude, or a coding agent about your product, it helps them route to your real pages instead of guessing from stale snippets. It lists canonical docs paths and summaries and prioritizes quickstarts, guides, and API references.
Every DocsAlot project gets a hosted MCP endpoint at mcp.docsalot.dev/yourco. The MCP server reads from the same maintained docs you publish for humans and exposes search_docs for retrieval, get_page for exact source text, and run_example for guided examples and snippets — without your team wiring infrastructure.
Yes. Team and Enterprise customers can host on a subdomain of their choice, such as docs.yourco.com, with automatic SSL. Agent-facing outputs stay attached to the same published docs layer.
No. A CMS publishes pages. DocsAlot publishes a full AI onboarding surface: human-readable docs, llms.txt, skill.md, hosted MCP tools, and audit signals that show where AI answers drift.
Yes. The done-for-you Docs Pack can import support articles, internal notes, README files, raw Markdown, Confluence spaces, and Google Docs into a clean DocsAlot tree, normalized into a navigable structure and reviewed with your team before launch.
Your domain is run through internal checks: whether it has llms.txt, whether the IA is agent-legible, whether MCP-style retrieval would work, and which high-intent prompts your docs currently miss. You get a 4–6 page report with practical fixes and citation-gap recommendations you can implement with or without DocsAlot.
Startup and Team plans have no per-seat charge for collaborators. Enterprise includes unlimited seats.
Pricing starts at $79/month for the Startup plan. The pricing page lists a free AI visibility audit as a standalone offering, but no free plan or trial for the platform is described.