Personal RAG
your own ChatGPT trained on your docs
// official site: quivr.com ↗
AI / LLM · PRO TIER
Quivrpro
Quivr is open-source RAG knowledge base — ingest documents (PDF, DOCX, video, audio, code repos), chat with them, build topic-specific "second brains". Modern alternative to AnythingLLM with stronger multi-modal ingestion.
// What it is
A closer look.
Quivr is open-source RAG knowledge base — ingest documents (PDF, DOCX, video, audio, code repos), chat with them, build topic-specific "second brains". Modern alternative to AnythingLLM with stronger multi-modal ingestion.
// Use cases
What it's for.
Concrete scenarios where teams pick Quivr over the SaaS alternative.
Team knowledge base
shared brains per project / topic
Multi-format ingestion
beyond PDFs (audio, video, code repos)
Research assistant
chat with academic libraries
Customer support AI
train on product docs + FAQs
// Who it's for
Built for these teams.
If your team profile matches one of these, Quivr is a strong fit out of the box.
Knowledge workers
building searchable personal libraries
Research teams
chatting with paper collections
Customer support teams
building AI-assisted help
Educators
providing students AI access to course material
Indie SaaS
offering "chat with your docs" features
// Differentiators
Why teams pick Quivr.
When evaluating self-hosted options for this category, here are the dimensions on which Quivr consistently lands above the alternatives.
- ✓Apache 2.0 — fully open
- ✓Multi-modal — handles audio + video + text + code
- ✓Multi-brain architecture — separate knowledge contexts
- ✓Modern UI — designer-quality interface
- ✓Active development — backed by Quivr (FR) — European OSS
- ✓Strong roadmap — frequent feature additions
// Integrations
Connects to.
The stack you'll plug Quivr into — services, protocols, and adjacent apps in the BluixApps catalog.
◇
LLM providers
OpenAI, Anthropic, Ollama, custom
◈
Document loaders
PDF, DOCX, Markdown, audio (Whisper), video, web pages, code repos
◆
Vector store
Supabase pgvector (default) or PostgreSQL+pgvector
▣
Auth
Supabase Auth (OAuth + email)
▦
Embeddings
OpenAI, Ollama, sentence-transformers
▩
Frontend
Next.js with API for custom integrations
▼
MCP
emerging tool protocol support
// Adoption & deployment
Notable users & community
- 38k+ GitHub stars (rapidly growing)
- Active Discord
- Backed by Quivr SAS (FR) — European OSS company
- Featured in RAG platform comparisons
- Strong release cadence with breaking changes between versions
What we ship
- Quivr official docker-compose (multi-service)
- Persistent volumes for Postgres + Quivr data
- Port 3002 exposed
- Pre-configured to detect Ollama on same VPS
- Compose at
/opt/quivr/repo/docker-compose*.yml - HTTPS via Let's Encrypt reverse proxy
- Backup hook covers Postgres + uploaded docs
// Tips & operations
Run it properly.
Operational guidance from running this in production — what to lock down, what surprises people.
// PERFORMANCE
Complex multi-service stack
4 GB RAM minimum, 8+ for production
// SECURITY
First-run wait
5-10 min for full startup
// OPERATIONS
Breaking changes
Quivr evolves fast; pin versions for stability
// RELIABILITY
Pair with Ollama
for local LLM + embeddings
// DEPLOYMENT
Backup priority
knowledge bases = sunk effort, irreplaceable
// SCALING
API token security
admin API has full access; rotate frequently
2048
// min ram (MB)
8
// min disk (GB)
8888
// access port
http
// protocol
pro
// bluixapps tier
// Alternatives in AI / LLM
Compare with
Project resources
Official sitequivr.com ↗