For the complete documentation index, see llms.txt.
Skip to main content

Turn your APIs AI-Native, without rebuilding them.

reShapr is the open source runtime that turns the REST, GraphQL, and gRPC APIs you already have into governed, LLM-optimized MCP tools. No bespoke MCP server to build and maintain for every API.

Apache-2.0 · Model Context Protocol · Stable 1.x (semver)

Your APIs weren't designed for agents. reShapr reshapes them.

Agents need small, precise, governed actions. Your APIs return large, general-purpose responses. reShapr closes that gap.

Context overload

Raw API responses waste tokens and degrade tool selection. reShapr filters outputs and composes calls so the model only sees what matters.

MCP server sprawl

One hand-built MCP server per API and per team means one security review and one cost footprint each. reShapr gives you a single governed control plane.

Non-deterministic workflows

Multi-step workflows get re-derived on every run. Custom Tools turn them into one compact, governed action.

The feedback we received

The first results from our reShapr evaluation at AXA are very promising. It addresses complex enterprise challenges with rare clarity and a true open source heart. As a maintainer in the CNCF space, I’m excited to support this community effort and see how it transforms how we expose services to agents.
Sebastien DegodezSoftware Engineer | TechLead
When organizations have hundreds or even thousands of APIs to maintain, speed and security become non-negotiable. reShapr transforms how existing services are exposed to agents, delivering an unmatched FinOps ratio and the fastest path to production. It’s a definitive solution for scaling agentic context within highly regulated environments.
Ludovic PourratAPI Architect | Platform Architect
I’m excited by reShapr’s clear purpose and the co-founders’ vision, which directly address the real-world challenges I see across the organizations I work with. Open-sourcing such a critical component to advance MCP democratization is a meaningful step forward for the ecosystem, and I’m genuinely looking forward to being part of this journey and helping build a strong, open community around reShapr.
Sébastien BlancJava Champion | Technical Director
I use reShapr to let AI assistants work with business systems: PayCaptain’s payroll API and, for a client in fashion retail, a PLM product catalog they can query in plain language and sync to multiple Shopify stores (products and prices), served from one consolidated MCP server built with cross-service scripted Custom Tools.

From API contract to agent-ready tools in three steps

1. Import

Point reShapr at an OpenAPI, GraphQL, or Protobuf/gRPC contract. Working MCP tools are generated immediately.

2. Reshape

Rename and pin arguments, orchestrate several services into one tool, filter responses, and choose tools, code, or hybrid exposure.

3. Govern and expose

Protect endpoints with API key or OAuth, set quotas, and version everything in Configuration Plans. Then run it on your own infrastructure.

Start building with reShapr

Follow the quickstart, import your first API contract, and give your agents governed MCP tools in minutes.

npm install -g @reshapr/reshapr-cli --allow-scripts=@scarf/scarf

What you get

A runtime that makes your existing APIs efficient, governed, and ready for any MCP client.

Context Control

Custom Tools reshape generated tools into the exact action an agent needs, and Tools Output Filtering trims payloads before they reach the model.

Code Mode

Let agents search tools, discover API types, and run code in a hardened sandbox instead of loading every tool definition into context.

Governance & security

API key or OAuth protection, backend secrets, operation filters, quotas, and audit, applied before any agent can call a tool.

Zero rewrite

Your backends stay untouched. Import an OpenAPI, GraphQL, or Protobuf contract and get working MCP tools for any MCP client.

Run anywhere

Use the CLI on your laptop, Docker Compose, or Kubernetes with Helm charts and controllers, all inside your own trust domain.

Open and neutral

Apache-2.0 licensed, with organization-balanced governance, built by the co-founders of Microcks.

Measured on a pull-request-velocity task against the official GitHub MCP server.

Same task, same data: the official GitHub MCP server needs 31 tool calls, reShapr needs one 👉 Read the benchmark.

What the user asks
Using the available GitHub MCP tools, analyze recent pull request velocity for the public repository microsoft/vscode.

Use the 10 most recently updated pull requests. For each pull request, collect the pull request details, up to 5 review touchpoints, and up to 10 changed file paths.

Return a compact summary with pull request id, title, creator, state, creation and merge timestamps, review touchpoints, and changed paths.
What reShapr exposes: one task-shaped tool, get_repo_velocity_metrics
Comparison of the official GitHub MCP and reShapr on the same task: 31 versus 1 agent tool calls, 34 versus 4 MCP HTTP roundtrips, 28,081 ms versus 2,082 ms tool-call latency, 267,824 versus 7,401 response bytes, and 66,956 versus 1,851 estimated tokens.Comparison of the official GitHub MCP and reShapr on the same task: 31 versus 1 agent tool calls, 34 versus 4 MCP HTTP roundtrips, 28,081 ms versus 2,082 ms tool-call latency, 267,824 versus 7,401 response bytes, and 66,956 versus 1,851 estimated tokens.

See also: a fashion brand runs its product catalog and Shopify syncs through one consolidated MCP server, built with cross-service Custom Tools 👉 Read the story.

Agent View