The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Tela
Open-source, self-hostable markdown team wiki with a built-in MCP server — so AI agents read, write, and search your docs as first-class teammates. Live multiplayer, semantic +…
Install
claude mcp add --transport http tela https://telawiki.com/api/mcpInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →The wiki that writes itself — from the code you already ship.
Home · Pricing · Compare · MCP for agents · Docs
tela is a self-hostable, markdown-native team wiki built for a world where agents are first-class authors and readers. It pairs a Go + PostgreSQL backend with a React 19 / Milkdown editor, live Yjs collaboration, ranked full-text and semantic search, and a built-in Model Context Protocol (MCP) server — so the same knowledge base your team edits in the browser is one your agents can search, read, and write directly. Atlas, its documentation engine, turns the artifacts you already produce into maintained wiki pages. Your content stays canonical markdown forever — pages.body is markdown, there is no proprietary block store.
👉 Start at telawiki.com — the marketing site walks through what tela does (with pricing, a comparison, and the agent/MCP story), and the hosted instance has a free tier: no install, no card.
Why tela
- Atlas auto-doc-gen — point Atlas at your sources and it drafts and maintains real wiki pages, so docs track what you ship instead of rotting.
- Built-in MCP server —
/api/mcpis part of the backend, not a bolt-on. Claude, Cursor, and other agents search, read, and author pages with scoped, per-tool write permissions. - Semantic + full-text search — ranked PostgreSQL FTS works out of the box; add an embedder for
pgvector-backed semantic retrieval and grounded "ask your docs" answers. - Live collaboration — real-time multi-cursor editing over Yjs in a Milkdown editor, with comments, backlinks, and revision history.
- Teams that run themselves — users create orgs and invite teammates by email, no admin ticket. Flip a space to public and it becomes a login-free blog surface with per-author home pages.
- Self-host & own your markdown —
pages.bodyis canonical markdown forever (no block table). Sync over WebDAV, export to zip/PDF, and run the whole stack with onemake up.
Quickstart
You need Docker (with Compose) and make. The bundled stack builds every image via Compose, so no host Node or Go toolchain is required.
git clone https://github.com/zcag/tela.git
cd tela
# 1. Write deploy/.env from the example with strong generated secrets
# (fills TELA_API_KEY_SECRET, TELA_SHARE_SECRET, TELA_PG_PASSWORD via openssl rand)
make setup
# 2. Edit deploy/.env — set TELA_PUBLIC_BASE_URL, and optionally
# TELA_ADMIN_* and TELA_SMTP_* (see Configuration below)
# 3. Build and start the full stack
make up
The stack comes up behind Caddy on http://localhost:8780. On first boot tela runs its embedded migrations automatically and lands you on the /setup wizard to create the first admin — unless you set TELA_ADMIN_PASSWORD in deploy/.env, which bootstraps the admin non-interactively.
One-click cloud deploy
Prefer managed hosting over running Compose? Deploy the published multi-arch images straight to a platform:
Similar plugs
Reverse engineer anything with agents, from app behavior down to native binaries.
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype…
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
Open-source web crawler and scraper for LLMs and AI agents: any website into clean, LLM-ready Markdown. Run it yourself, or use Crawl4AI Cloud with one key.