GenesisPod

An open-source, enterprise-grade platform for AI deep research, content production, and multi-agent collaboration.

English · 简体中文

---

What is GenesisPod?

GenesisPod is a full-stack platform for building and running AI research and multi-agent workflows in production. It ships a complete product surface — deep research, multi-agent orchestration, document/slide generation, long-form writing, RAG and knowledge graphs — on top of a strictly layered, architecture- governed backend.

It is fully open source under AGPL-3.0, with a commercial license available for closed-source and SaaS use.

Why another AI platform?

The differentiator is architectural discipline you can verify. Most agent frameworks rot into a tangle of cross-imports within a year. GenesisPod enforces its 5-layer boundaries (ai-app → ai-engine → ai-harness → ai-infra → open-api) through three independent gates — ESLint rules, a jest architecture-spec suite, and a pre-push + CI merge gate — so the structure stays sound as the codebase grows. The architecture compliance is machine-checked on every push, not a diagram in a wiki that drifts from reality.

Features

• AI Research — multi-step planning, source gathering, and report generation.
• Agent Playground — multi-agent mission orchestration with live tracing, token/cost accounting, and structured report artifacts.
• AI Ask / Insights — multi-model Q&A and topic insights.
• AI Office / Slides / Writing / Social — document, presentation, long-form, and social content generation.
• Library / RAG / Knowledge Graph — ingestion, retrieval, and knowledge consolidation.
• BYOK & multi-provider — OpenAI, Anthropic, Gemini, Grok, DeepSeek, and any LiteLLM-compatible provider, with a first-class secrets module.
• Admin — model, tool, secrets, data-source, and system management.

Architecture

A monorepo with three runtimes:

Package	Stack
frontend/	Next.js 14, React 18, TypeScript, Tailwind, Zustand, SWR, TanStack Query
backend/	NestJS 10, Prisma, PostgreSQL 16, Redis 7, Socket.IO
ai-service/	FastAPI (auxiliary AI service)

The backend is layered into five top-level modules with a strict one-way dependency direction:

open-api    →  external API / MCP / admin surface
ai-app      →  business applications (research, teams, office, writing, ...)
ai-engine   →  reusable AI primitives (LLM, tools, RAG, knowledge, planning)
ai-harness  →  multi-agent runtime, lifecycle, evaluation, protocols
ai-infra    →  auth, storage, secrets, notifications

ai-app reaches ai-engine only through a facade; ai-engine never imports ai-harness. See STRUCTURE.md for the full map.

Quick start

Requirements

• Node.js >= 20, npm >= 9
• Docker / Docker Compose
• At least one model provider API key (e.g. OPENAI_API_KEY)

Run it

# 1. Install
npm install

# 2. Configure — copy the template and fill in your keys (never commit .env)
cp .env.example .env

# 3. Start infrastructure (postgres + redis + flaresolverr)
npm run db:setup

# 4. Initialize the database
cd backend
npm run prisma:generate && npm run prisma:migrate && npm run prisma:seed
cd ..

# 5. Start the full stack
npm run dev

Default ports: frontend http://localhost:3000, backend http://localhost:3001, AI service http://localhost:5000.

Run a single side with npm run dev:frontend / dev:backend / dev:ai.

Minimum environment variables

DATABASE_URL, REDIS_URL, JWT_SECRET, and one provider key (OPENAI_API_KEY or another). For login, storage, and integrations see .env.example.

Development

Command	Purpose
npm run dev	full-stack dev
npm run type-check	TypeScript check
npm run test:quick	fast tests
npm run verify:arch	architecture-boundary checks
npm run verify:full	lint + type + test + build
npm run e2e	Playwright end-to-end

See CONTRIBUTING.md before opening a PR.

License

GenesisPod is dual-licensed:

• AGPL-3.0 for open-source and self-hosted use. Note: AGPL treats network use as distribution — if you run a modified version as a service, you must publish your source changes.
• Commercial license for closed-source products, proprietary SaaS, or when you need warranty/SLA/indemnity. Contact hello@gens.team.

Not sure which you need? See the decision table.

Contributing

Contributions are welcome. Please read CONTRIBUTING.md and note that a one-time CLA signature is required (automated via a bot on your first PR) — this is what keeps the dual-license model possible.

Security

Found a vulnerability? Do not open a public issue — see SECURITY.md.