Skip to content

Open Source Β· MIT Licensed

The AI Brain for Your Entire Team

Ingest your docs. Ask in natural language. Get cited answers via chat, voice, or WhatsApp. Self-host or use our cloud.

πŸ”“ MIT Licensed 🐳 4 Deploy Options πŸ”‘ Bring Your Own LLM Keys πŸ›‘οΈ Full Data Sovereignty
πŸ’¬ Try Raven β€” Ask anything Interactive Demo

How does Raven handle document ingestion?

Raven supports PDF, DOCX, Markdown, HTML, and web URLs. Documents are automatically chunked, embedded, and indexed using hybrid vector + BM25 search for accurate retrieval.

πŸ“Ž source: docs/ingestion.md, lines 14-28

Interactive demo β€” real AI answers coming soon

Your team's knowledge is scattered. Your people are stuck.

Docs nobody reads

Knowledge lives in PDFs, wikis, and Confluence pages that go stale. Your team asks the same questions on Slack instead.

Search that doesn't work

Keyword search returns 50 results. None of them answer the actual question. Your team gives up and pings someone senior.

No way to just ask

Your team wants to ask a question and get an answer β€” with a source. Not browse. Not search. Just ask.

Raven fixes this. Ingest your docs. Ask in natural language. Get cited answers β€” via chat, voice, or WhatsApp.

Three channels. One knowledge base.

Your team gets answers where they already work β€” chat, voice, or WhatsApp.

AI Chat

Ask your docs anything

Semantic search powered by hybrid vector + BM25 ranking. Every answer comes with source citations.

Raven AI Chat interface
Voice Agent

Talk to your knowledge base

Real-time voice calls via WebRTC. Hands-free access powered by LiveKit.

Raven Voice Agent UI
WhatsApp

Meet your team where they already are

Connect via WhatsApp Business API. Zero app installs, zero training.

Raven WhatsApp Conversation

Simple by design.

Three steps from scattered docs to instant answers.

1

Ingest

Upload PDFs, DOCX, Markdown, HTML, or paste URLs. Raven chunks, embeds, and indexes everything automatically.

Automatic chunking & embedding

2

Search

Hybrid retrieval β€” pgvector cosine similarity + BM25 full-text ranking fused with Reciprocal Rank Fusion. No keyword guessing.

Vector + BM25 hybrid ranking

3

Answer

Your team asks in natural language. Raven responds with accurate, cited answers β€” via chat, voice, or WhatsApp.

Cited answers across 3 channels

Your infrastructure. Your rules.

Deploy Raven anywhere β€” from a Raspberry Pi to a Kubernetes cluster.

🐳

Docker Compose

One command. Full stack. Perfect for getting started and small teams.

☸️

Kubernetes

Production-grade, multi-region, auto-scaling.

πŸ–₯️

EC2 + Ansible

Cloud VM with auto-TLS via Traefik. Playbooks ready to go.

πŸ“‘

Edge / Raspberry Pi

25MB ARM64 binary. Run the API on a Pi, connect a remote AI worker via gRPC.

Built for teams that don't compromise.

Enterprise-grade security, compliance, and control β€” out of the box.

Multi-Tenant Isolation

Full Row-Level Security per tenant in PostgreSQL. Your data never crosses boundaries.

SSO / SCIM

Keycloak-powered OIDC and SAML out of the box. SCIM coming soon

Audit Logs Coming soon

Timestamped audit trail for every action. Compliance-ready.

BYOK

Bring your own LLM keys β€” Anthropic, OpenAI, Cohere, or self-hosted models. Zero vendor lock-in.

eBPF WAF

Kernel-level XDP pre-filtering and observability at the edge.

GDPR / SOC2

Built for compliance. Self-host for full data sovereignty.

Simple pricing. Generous free tier.

Start free, scale when you're ready. Every plan includes the full open-source platform.

Free

$0 forever

Everything you need to get started with AI-powered knowledge.

  • AI chat, voice & WhatsApp
  • Bring Your Own Key (BYOK)
  • Self-hostable
  • Unlimited documents
  • Full MIT source

Limits

  • 5 users
  • 2 workspaces
  • 3 knowledge bases
  • 500 MB storage
  • 1 voice session
  • 60 min/mo voice
⭐ Recommended

Pro

$29 /month

For growing teams that need higher limits and priority support.

  • Everything in Free
  • Higher limits
  • Advanced rate limiting
  • Priority support

Limits

  • 25 users
  • 10 workspaces
  • 50 knowledge bases
  • 10 GB storage
  • 5 voice sessions
  • 1,200 min/mo voice

Enterprise

Contact Sales

For organizations that need SSO, audit logs, and dedicated support.

  • Everything in Pro
  • SAML / SSO / SCIM
  • Audit logs
  • eBPF WAF
  • Dedicated support
  • Custom SLA

Limits

  • Unlimited users
  • Unlimited workspaces
  • Unlimited knowledge bases
  • Unlimited storage
  • Unlimited voice sessions
  • Unlimited minutes

Start building with Raven in under 5 minutes.

docker compose up -d