MySearch-Proxy
MySearch-Proxy is a unified search backend for AI agents, bundling an MCP server and skills plus a deployable proxy/console. It routes queries to upstream providers (Tavily, Firecrawl, Exa, optional X/Social), manages provider health and auth errors, provides token/key/quotas syncing via the proxy layer, and offers MCP tools (search, extract_url, research, mysearch_health) accessible over stdio/HTTP/SSE.
Score Breakdown
⚙ Agent Friendliness
🔒 Security
Strengths: proxy introduces an admin password concept and surfaces upstream auth failures via health checks (auth_error/live_status/live_error), which can reduce accidental use with invalid credentials. Concerns/unknowns: TLS enforcement, least-privilege scopes, and secret-handling guarantees (e.g., avoiding logging secrets) are not explicitly documented in the provided README. Scope granularity is likely coarse because the README focuses on API keys/tokens and upstream key pools rather than fine-grained scopes.
⚡ Reliability
Best When
You want to plug an MCP-based search capability into AI tooling while centralizing provider configuration/health checks behind a proxy service.
Avoid When
You cannot tolerate dependency on multiple third-party search providers (Tavily/Firecrawl/Exa/X/Social) or you require strict guarantees about data residency/compliance, which are not described here.
Use Cases
- • Single unified search tool for Codex/Claude Code/OpenClaw and other agents via MCP/skills
- • Team deployment to share a common search backend and centrally manage upstream provider keys/tokens/quotas
- • Domain-aware document/web search using provider-specific routing and fallback behavior (e.g., Firecrawl domain-first then Tavily)
- • Extracting and researching URLs (search + fetch/body extraction) with health/status visibility
- • Compatibility endpoint for social/search (/social/search) and X/Social fallbacks
Not For
- • Use as a general-purpose crawler without respecting provider terms and auth
- • Environments requiring formally verified security/compliance guarantees without further audit
- • Applications needing an OpenAPI/REST contract or SDKs specifically documented for third-party developers (not clearly provided in supplied README)
Interface
Authentication
Authentication model appears to be proxy console/admin password plus API keys for client access (and upstream provider keys managed/validated by the proxy). No OAuth/scopes are described in the provided README.
Pricing
Pricing for this project itself is not described; it functions as a gateway that consumes upstream provider APIs (Tavily/Firecrawl/Exa/etc.), which will drive costs.
Agent Metadata
Known Gotchas
- ⚠ Caching is TTL-based and in-process; single CLI invocations may not benefit across processes (explicitly noted).
- ⚠ If upstream provider auth fails, routing behavior changes: routes/docs/resource skip providers marked with auth_error.
- ⚠ Firecrawl 'site:' empty-results triggers fallback logic; agents expecting strictly site-filtered results may see broader follow-up retrieval.
Alternatives
Full Evaluation Report
Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for MySearch-Proxy.
AI-powered analysis · PDF + markdown · Delivered within 30 minutes
Package Brief
Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.
Delivered within 10 minutes
Score Monitoring
Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.
Continuous monitoring
Scores are editorial opinions as of 2026-03-30.