qurio

Qurio is a self-hosted ingestion and retrieval (RAG) engine for AI coding assistants. It crawls or ingests local documents (e.g., web pages, PDFs, Markdown), chunks content structurally, embeds it (Gemini for embeddings), stores vectors/metadata in Weaviate/PostgreSQL, and exposes retrieval to agents via an MCP server over a JSON-RPC 2.0 endpoint.

Evaluated Mar 30, 2026 (22d ago)
Repo ↗ Ai Ml rag self-hosted mcp model-context-protocol vector-search hybrid-search weaviate postgresql document-ingestion code-assistants docker-compose
⚙ Agent Friendliness
49
/ 100
Can an agent use this?
🔒 Security
23
/ 100
Is it safe for agents?
⚡ Reliability
25
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
70
Documentation
65
Error Messages
0
Auth Simplicity
55
Rate Limits
5

🔒 Security

TLS Enforcement
20
Auth Strength
15
Scope Granularity
0
Dep. Hygiene
40
Secret Handling
45

Runs as a local service behind Docker; README does not specify TLS usage, MCP authentication/authorization, or how requests are protected. API keys (Gemini, optional rerank) must be provided via environment/settings; no guarantees are stated about secret handling in logs. Dependency hygiene and vulnerability posture cannot be verified from provided content.

⚡ Reliability

Uptime/SLA
0
Version Stability
40
Breaking Changes
30
Error Recovery
30
AF Security Reliability

Best When

You want localhost RAG for coding assistants and can run Docker Compose locally while allowing embedding calls to the configured provider (Gemini) and connecting your agent via MCP.

Avoid When

You need an SDK/OpenAPI-described REST API, strong documented rate limiting, or you cannot provide/secure required API keys (Gemini, optional rerank providers).

Use Cases

  • Grounded documentation search for local development assistants
  • Uploading/crawling internal engineering docs to reduce hallucinations in code generation
  • Providing an MCP tool interface (search/list/read) to agentic coding workflows
  • Hybrid keyword+vector retrieval with optional reranking

Not For

  • Production deployments that require strict enterprise security controls without additional hardening
  • Environments where outbound embedding calls (e.g., Gemini) are not allowed
  • Use cases needing fine-grained section-by-section retrieval (roadmap item)

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

OAuth: No Scopes: No

README describes adding API keys for Gemini and optional rerank providers, but does not describe any authentication/authorization mechanism for accessing the MCP endpoint itself. Treat the MCP endpoint as potentially unauthenticated unless configured otherwise in the code/deployment.

Pricing

Free tier: No
Requires CC: No

Open source (MIT) self-hosted; cost depends on external embedding/reranking providers (e.g., Gemini) and infrastructure (Weaviate/Postgres/Docker). No pricing for the service itself is described.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • MCP tools are described at a high level; exact tool argument schemas/response shapes and pagination/limit behavior are not documented in the provided README.
  • MCP transport is described as 'stateless, streamable HTTP'; agents may need HTTP MCP client support.
  • Indexing/ingestion is asynchronous; agents may query before ingestion completes unless the user coordinates workflow via the dashboard/status.

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for qurio.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered