Quivr hosting

A second brain you can ask questions — files, links and notes in one place.

Agents & RAGApache-2.0upstream discontinuedfrom $11.99/mo

In short

Quivr hosting at AIHostBox means a dedicated, single-tenant Quivr instance on our infrastructure: we install it, pin it to a reviewed release, put it behind TLS on your domain, back it up nightly and patch everything around it. Plans start at $11.99 per month and the admin account, the data and any API keys are yours.

No longer maintained upstream

The Quivr web app has had no release since v0.0.322 in October 2024; the project now develops only the quivr-core Python library.

We still host it for teams who depend on it: we keep the instance running, backed up and behind TLS, and patch the operating system and the services around it. We cannot fix bugs or vulnerabilities inside Quivr itself, and there will be no feature updates. Order it only if you accept that; if you are starting fresh, the alternatives below are the safer choice.

BrainsMulti-formatSharingMulti-model

Quivr began as an open-source “second brain”: you upload PDFs, Office files, spreadsheets, audio and web links into brains, then chat with each brain on its own terms. Brains can be private or shared with a link, and each can point at a different model, including local ones through Ollama. The project is maintained by the team behind Quivr.com, now published under The Vibe Company on GitHub, and is licensed Apache 2.0.

The web app is more than a single container. It is a Next.js front-end, a FastAPI backend, Celery workers for parsing and crawling, Redis as the queue, and Supabase for PostgreSQL with pgvector, authentication and file storage. Keeping that many moving parts healthy is what makes Quivr awkward on a laptop, and why it is easier to use when someone else runs the stack and you just open a browser tab.

Quivr hosting plans

PlusRecommended minimum

4 GB RAM · 2 vCPU · NVMe

$11.99/mo

Order

ProTeams & heavier load

8 GB RAM · 3 vCPU · NVMe

$18.99/mo

Order

Every plan is a dedicated single-tenant instance with NVMe storage, nightly off-box backups, TLS and managed patching. Quarterly, semi-annual and annual terms are discounted at checkout. Model tokens are not included — you bring your own key.

How we run Quivr

We deploy the last full-stack Quivr release (v0.0.322) from its official docker-compose: the Next.js front-end on port 3000, the FastAPI backend-api, Celery worker, beat and notifier containers, Redis, and a Supabase stack providing PostgreSQL with pgvector, auth and storage.

What we configure for you

OPENAI_API_KEY or your ANTHROPIC_API_KEY, MISTRAL_API_KEY or GROQ_API_KEY
OLLAMA_API_BASE_URL if you want a self-hosted model instead
A fresh JWT_SECRET_KEY and Supabase keys in place of the published demo values
QUIVR_DOMAIN, BACKEND_URL and NEXT_PUBLIC_BACKEND_URL set to your domain over TLS
TELEMETRY_ENABLED off and DEACTIVATE_STRIPE on for a private install
QUIVR_SMTP_SERVER and related settings so invitations and resets are delivered

Which plan fits

Plus (4 GB) runs the whole stack for one person or a couple of users with modest document sets. Pro (8 GB) gives the Supabase database and the parsing worker room for a small team uploading larger PDFs and spreadsheets at the same time.

The honest limits

The Quivr web app is no longer released upstream, so we patch the operating system and containers but you should not expect new Quivr features. If you want an actively developed alternative for the same job, AnythingLLM or Open WebUI’s knowledge features are closer to where the ecosystem is going.

What people run Quivr for

Personal knowledge base

One place for the things you keep re-reading.

Small team wiki with answers

Shared brains with per-brain models.

Per-client document sets

Give each client or project its own brain so answers never mix sources.

Licence and version policy

Item	Position
Upstream licence	Apache-2.0
Sold as managed hosting	Yes
Tenancy	Single tenant — one customer per instance
Patching	Critical advisories within 24 hours, high severity within 72

Our full position on licences, including the applications we decline to sell, is on the licensing page.

Works well with Quivr

More Agents & RAG we host

Questions

Quivr hosting: your questions

Is Quivr still being developed?

The upstream repository now focuses on quivr-core, a Python RAG library, and the full web app’s last tagged release was v0.0.322 in October 2024. We host that final release of the app. It works as described, but new features are going into the library rather than the app you log into.

Which models can I use?

The app works with OpenAI by default, and its configuration also accepts Anthropic, Mistral and Groq keys. For a self-hosted model, set OLLAMA_API_BASE_URL to an Ollama instance. Embeddings are configured separately, and switching embedding models after you have uploaded documents means re-uploading them.

What file types can I add to a brain?

Text, Markdown, PDF, PowerPoint, Excel, CSV, Word, audio and video files, plus web links that the crawler fetches. Parsing runs in the background Celery worker, so large uploads appear in the brain a little after the upload finishes rather than instantly.

How do logins work?

Authentication is handled by Supabase. Email and password is the default mode, set with NEXT_PUBLIC_AUTH_MODES. We replace the sample admin account and the demo JWT secret from the example configuration before handing the instance over, and you invite colleagues from there.

What exactly gets backed up?

Everything lives in Supabase: the PostgreSQL database with brains, chats and pgvector embeddings, plus Supabase storage for the original files. Both are part of the nightly off-box backup. Redis only holds the job queue, so nothing is lost if it restarts.

Run Quivr without running a server

A dedicated Quivr instance, deployed and maintained by us. Month to month, cancel whenever.

Browse the catalogue Talk to a human