What Quivr does, how to run it, and how to send it content to search and monitor. For integrators, operators and non-developers.
- Security policy: report vulnerabilities privately and verify release inventories
- Public HTTP transport contract: how the HTTP API contract is organised and checked
- docs-site/run-quivr/scale-quivr.mdx: capacity budgets, import lessons and incident checks for operators
- Quivr V2 — Remaining limits: known limits and what is not claimed
- Size PostgreSQL for imports: size PostgreSQL connections and memory for imports
- Shared infrastructure settings: resolve shared profiles, apply deployment settings and check drift
- Deploy a search measurement store: deploy an authenticated shared search measurement store
- Serve EmbeddingGemma 2 on Modal: authenticated Modal embedding deployment, corpus rebuild and rollback
- Railway evaluation demo: run a hosted single-node evaluation demo
- Scale bulk workers on Railway: scale Railway bulk workers from queue backlog
- Switch hosted text embeddings with rollback: evaluate, fill and switch hosted text embeddings with rollback
- docs-site/run-quivr/dedicated-vector-host.mdx: provision an authenticated index VM and rebuild it from stored content
- docs-site/run-quivr/single-machine.mdx: run a complete CPU evaluation stack from one Compose release bundle
- Quivr Search demo (THE-663): the demo web app: search, sources, live feed and the admin view
- Command-line reference: every
quivrcommand, what it needs to run, its flags and exit codes - HTTP API reference: every endpoint, field and example, generated from the contract
- MCP reference: every MCP profile and tool an AI agent can use, with its arguments
- Quivr V2: the project overview: why Quivr, architecture, quickstart and what works today