80+ CLI commands · 134 API endpoints

Learn and apply — knowledge that works

Auto-extract knowledge cards from documents, equip AI agents with skills, and let them work across every page. Self-hosted, API-first, fully under your control.

Everything you need to manage knowledge

A complete platform for documents, entries, AI agents, and automation.

Documents

Upload PDF, DOCX, Markdown, and more. Automatic text extraction, vector embeddings, and semantic search powered by pgvector.

Entries

Rich wiki entries with a Notion-style editor. AI-assisted content generation, categories, tags, comments, and full-text search.

Cards

Knowledge cards extracted from documents via AI. Capture key insights as structured, searchable, and taggable units.

Skills

Agent skill system with file management. Create, clone, export/import as ZIP, and share skills across your organization.

Collections

Smart collections with query-based auto-population. Manage member and follower roles with fine-grained visibility controls.

CLI & API

80+ CLI commands with JSON-only output, built for AI agents. 134 REST API endpoints with API key authentication.

Built for automation

A non-interactive CLI that outputs pure JSON. Designed for AI agents and scripts.

terminal

# Install Wiki CLI (Linux & macOS)

$ curl -fsSL https://www.wiki.seeyon.chat/downloads/install.sh | bash

# Verify installation

$ wiki --version

# Login with your API key (get from web UI: Account Settings → API Key)

$ wiki login --save-api-key=wiki_xxxxxxxx_yyyyyyyyyyyyyyyy

# Verify authentication

$ wiki whoami --verify

80+

CLI commands

134

API endpoints

JSON

Output format, always

Agent-native architecture

AI agents are first-class citizens. Every page, every workflow, every API is designed with agent consumption in mind.

Page Agents

Every list and detail page has a dedicated AI agent with context-aware tool calling. Ask questions, get summaries, or modify content.

Task Workspaces

Pre-agent workspace builder packages documents, entries, and data snapshots into isolated environments for complex agent tasks.

Diff Proposals

Agents propose edits as structured diffs. Review changes in a real diff viewer before approving or rejecting modifications.

AI Extraction

Extract knowledge cards, generate summaries, suggest tags, and create wiki entries from uploaded documents automatically.

agent workflow

1. User asks question on document page

2. Page agent receives context + tools

3. Agent calls semantic_search, read_document

4. Agent proposes edit as structured diff

5. User reviews diff → approve / reject

6. Changes applied, chat persisted to DB

Deploy anywhere

Self-host with Docker, connect your own database and LLM. No vendor lock-in.

Self-hostable

Run on your own infrastructure with Docker. Full control over your data and models.

Multi-tenant

Hostname-based organization isolation. Each tenant gets its own database and storage.

API key auth

Personal API keys for service-to-service access. JWT cookies for browser sessions.

Bring your own LLM

Compatible with any OpenAI-compatible endpoint. Use local models or cloud providers.

docker-compose.yml

# docker-compose.yml
services:
  wiki:
    image: seeyonai/wiki:latest
    ports:
      - "4320:4320"
    environment:
      DATABASE_URL: postgresql://wiki:secret@db:5432/wiki
      JWT_SECRET: your-secret-key
      OPENAI_API_KEY: sk-...
      OPENAI_BASE_URL: https://api.openai.com/v1
      KB_STORAGE_ROOT: /var/app/wiki-data
    volumes:
      - wiki-data:/var/app/wiki-data

  db:
    image: pgvector/pgvector:pg16
    environment:
      POSTGRES_DB: wiki
      POSTGRES_PASSWORD: secret