openrouter-mcp-multimodal

A Model Context Protocol (MCP) server that routes requests to OpenRouter models and exposes tools for text chat (including multimodal message content), image analysis (with server-side image fetching and Sharp-based optimization), image generation, and model search/validation.

Evaluated Mar 30, 2026 (33d ago)
Homepage ↗ Repo ↗ Ai Ml mcp openrouter multimodal vision image-generation sharp typescript nodejs ai-tools
⚙ Agent Friendliness
71
/ 100
Can an agent use this?
🔒 Security
58
/ 100
Is it safe for agents?
⚡ Reliability
40
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
85
Documentation
80
Error Messages
--
Auth Simplicity
95
Rate Limits
30

🔒 Security

TLS Enforcement
70
Auth Strength
70
Scope Granularity
20
Dep. Hygiene
60
Secret Handling
70

README describes SSRF mitigations for URL fetching (blocked private/link-local/reserved targets, capped response size, timeouts, redirect cap) and warns about local file access and disk writes via save_path. It does not document TLS enforcement details or explicit secret logging behavior, but it uses env var configuration for API keys.

⚡ Reliability

Uptime/SLA
0
Version Stability
65
Breaking Changes
40
Error Recovery
55
AF Security Reliability

Best When

You want a ready-to-use MCP server to unify OpenRouter text + vision + image generation behind a small tool surface in common MCP clients.

Avoid When

You need strict control over where image bytes come from (local filesystem and URL fetching are both possible) or you require fully specified operational guarantees (SLA, explicit retry/idempotency semantics, rate-limit headers) beyond what’s documented.

Use Cases

  • Multimodal chat with OpenRouter models via MCP clients (e.g., Claude Desktop, Cursor)
  • Vision workflows: analyze user-provided images from local paths, URLs, or data URIs
  • Image generation from text prompts via OpenRouter
  • Model discovery: search/validate models and retrieve model info (pricing/context/capabilities)

Not For

  • Running untrusted prompts that may cause the server to read unintended local files (tools can access local file paths provided by the MCP client)
  • SSRF-safe internal network access (the README notes SSRF mitigations, but you should still avoid exposing sensitive internal hosts)
  • Production environments that require explicit, published SLAs or strict compliance guarantees (not evidenced in the provided docs)

Interface

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

Authentication

Methods: OpenRouter API key via environment variable OPENROUTER_API_KEY (passed to MCP server process)
OAuth: No Scopes: No

Auth is delegated to OpenRouter using an API key provided to the MCP server runtime; no OAuth flow or fine-grained scopes are described in the README.

Pricing

Free tier: Yes
Requires CC: No

The README indicates support for free models on OpenRouter (not the MCP server itself). Exact rate/usage limits and costs are not specified here.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • `analyze_image` may access local files if the MCP client supplies paths; ensure the agent only passes intended, safe paths
  • `analyze_image` may fetch image URLs server-side; ensure the agent avoids URLs pointing to internal/sensitive hosts even though private IPs and response sizes/redirects are bounded
  • `generate_image` can write files when `save_path` is used; treat prompt and paths as untrusted input to avoid writing to unintended locations
  • Model search/validation tools may call OpenRouter endpoints; caching exists (TTL env var), but cache behavior and invalidation guarantees are not fully specified

Alternatives

Full Evaluation Report

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

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