{"id":"andyshaman-gemini-webapi-mcp","name":"gemini-webapi-mcp","homepage":null,"repo_url":"https://github.com/AndyShaman/gemini-webapi-mcp","category":"ai-ml","subcategories":[],"tags":["mcp","model-context-protocol","python","ai","gemini","image-generation","image-editing","file-analysis","browser-cookies","automation"],"what_it_does":"Provides an MCP server that bridges to Google Gemini for image generation/editing, file analysis (e.g., images/PDFs/videos), and text chat by using Gemini web access authenticated via browser cookies from Chrome (or manually provided cookie values).","use_cases":["MCP-enabled assistants that need Gemini-backed image generation/editing","Agents that analyze user-provided files (PDFs, documents, images, videos)","Chat and multi-turn conversations with Gemini through MCP clients (e.g., Claude Code/Desktop)","Local, developer-run automation that wants to avoid separate API keys by reusing an existing logged-in browser session"],"not_for":["Production-grade deployments requiring strong, first-class authentication/authorization controls","Teams that cannot store or process sensitive browser cookies","Environments needing guaranteed compliance/SLA, predictable uptime, and stable external API contracts","Use cases where watermark removal is disallowed or violates policies/laws"],"best_when":"You control the runtime environment (local machine/dev box), can provide/manage Chrome cookies safely, and want an MCP interface for Gemini capabilities quickly.","avoid_when":"You need enterprise-grade security boundaries, audited access control, or you cannot accept dependence on reverse-engineered web interactions and third-party session cookies.","alternatives":["Official Google Gemini APIs (via Google AI SDKs / REST)","Other MCP adapters that use official APIs or dedicated service accounts","Workflow tools that call Gemini via supported server-side APIs (with API keys)","If you only need image generation: direct vendor SDK/API rather than cookie-based web automation"],"af_score":48.0,"security_score":42.8,"reliability_score":25.0,"package_type":"mcp_server","discovery_source":["github"],"priority":"high","status":"evaluated","version_evaluated":null,"last_evaluated":"2026-03-30T15:30:54.495512+00:00","interface":{"has_rest_api":false,"has_graphql":false,"has_grpc":false,"has_mcp_server":true,"mcp_server_url":null,"has_sdk":false,"sdk_languages":[],"openapi_spec_url":null,"webhooks":false},"auth":{"methods":["Chrome cookie-based authentication via browser-cookie3","Manual cookie injection via environment variables (GEMINI_PSID, GEMINI_PSIDTS)","Account selection via GEMINI_ACCOUNT_INDEX"],"oauth":false,"scopes":false,"notes":"Authentication is not an API-key/API-token mechanism; it reuses browser session cookies (or manually provided cookie values). This implies sensitive credential material handling and dependency on Google web session behavior."},"pricing":{"model":"Google Gemini web experience (no direct API pricin","free_tier_exists":false,"free_tier_limits":null,"paid_tiers":[],"requires_credit_card":false,"estimated_workload_costs":null,"notes":"Project claims 'No API keys. Free' but actual usage limits/costs are determined by the underlying Google account/session and any rate/quotas on the web platform."},"requirements":{"requires_signup":true,"requires_credit_card":false,"domain_verification":false,"data_residency":[],"compliance":[],"min_contract":null},"agent_readiness":{"af_score":48.0,"security_score":42.8,"reliability_score":25.0,"mcp_server_quality":78.0,"documentation_accuracy":70.0,"error_message_quality":0.0,"error_message_notes":null,"auth_complexity":35.0,"rate_limit_clarity":10.0,"tls_enforcement":100.0,"auth_strength":30.0,"scope_granularity":10.0,"dependency_hygiene":55.0,"secret_handling":25.0,"security_notes":"Security posture is constrained by the reliance on sensitive session cookies (GEMINI_PSID/GEMINI_PSIDTS) and cookie extraction from a local browser. The approach lacks fine-grained scoping/authorization controls typical of API keys/OAuth. TLS enforcement for network transport is not explicitly documented, but typical HTTPS usage for web interactions is assumed; nonetheless, cookie handling and logging hygiene are critical. Dependencies include curl-cffi, numpy, Pillow, and browser-cookie3; specific CVE status and pinning are not shown in the provided data.","uptime_documented":0.0,"version_stability":30.0,"breaking_changes_history":30.0,"error_recovery":40.0,"idempotency_support":"false","idempotency_notes":null,"pagination_style":"none","retry_guidance_documented":false,"known_agent_gotchas":["Cookie-based auth may break when Gemini web session changes or cookies expire; agents may need to re-run auth/setup or use a 'reset' tool","Rate limiting/quotas from the underlying Gemini web endpoints are not clearly documented in the provided README; retries may trigger further failures","MCP tool behavior depends on the upstream web workflow; long-running image generation/upscaling may time out and fall back to native resolution","Watermark removal is an additional transformation that may be slow or fail depending on image properties and upstream output format"]}}