Your company knowledge,
one MCP URL away.
Gigamcp turns your GitHub, Google Drive, and other tools into a single hosted MCP server. Drop the URL into Cursor or Claude and your whole team queries permission-aware company context — no agent framework, no glue code.
No credit card. EU data residency. Bring-your-own connectors.
Built for teams, not demos.
Multi-tenant by design, with the boring-but-essential primitives — RLS, audit log, audience filters — wired up before day one.
Per-tenant connectors
GitHub, Google Drive, and the rest plug in once per workspace and respect your organisation chart.
Group-level ACL
Audience filters every retrieval call. Engineering can't see HR documents, even by accident.
Self-service onboarding
Sign up, paste an MCP URL into Cursor, and your team is querying company knowledge in 5 minutes.
Simple, per-seat pricing.
Start free. Upgrade when your team needs it.
Free
Try Gigamcp on a side project. No credit card.
StarterRecommended
For teams getting their first MCP-native workflows live.
Frequently asked questions
Does Gigamcp store my source data?⌄
Indexed embeddings live in your tenant partition of OpenSearch Serverless. Raw files stream from the source system on every read; we only persist what we have to.
How is access controlled?⌄
Every search and tool call carries a server-side audience filter (tenant + groups + user). The OpenSearch query-time filter is mandatory and applied in middleware, not the client.
Where is data stored?⌄
AWS Frankfurt (eu-central-1). EU GDPR-compliant by default. Custom regions available on the Enterprise plan.
Which MCP clients are supported?⌄
Anything that speaks the Streamable HTTP MCP transport — Cursor, Claude Desktop, VS Code Continue, and most OSS MCP clients.