ProtoMap is a context engine that maps gRPC services, Protobuf messages, and field-level dataflows to inform developers and coding agents over MCP.
What the product does and how it is positioned
ProtoMap is a context engine and model harness designed to map gRPC services and Protobuf schemas across distributed codebases. It performs static analysis on source code to catalog service callers, call sites, dependencies, and field-level dataflows across multiple repositories.
The platform supports both human developers and automated coding agents. Engineers can query service relationships in plain English through a web platform, while AI coding assistants interact with the map over Model Context Protocol to analyze architectural dependencies without manual multi-repository grepping.
Source-supported ways to use the product
Engineering teams map service dependencies up to five hops out to verify what will break before altering or deprecating gRPC methods and fields.
Autonomous coding tools access cross-repository dependency graphs via an MCP server to retrieve exact call-site relationships during refactoring tasks.
The documented workflow, where available
Verify host prerequisites including Git, Docker with Compose 2.23.1 or newer, 4 GB of RAM, and 20 GB of disk space.
Clone the ProtoMap repository and run the setup script to configure local services and the command-line interface.
Access the generated local address, submit the setup key, and create the organization and owner credentials.
Run the command-line scanning tool across target code repositories containing gRPC and Protobuf definitions.
Connect coding agents to the ProtoMap MCP server or use the web platform to query service relationships in natural language.
ProtoMap uses static analysis scanners to inspect source code and construct a deterministic map of gRPC services and Protobuf message structures. Instead of relying solely on runtime network tracing or high-level API specification files, the engine indexes exact call sites and field-level dataflows directly from repositories at the commit level.
By decoupling code discovery from generative reasoning, the engine provides structured graphs directly to downstream language models and coding agents. Agents query the Model Context Protocol endpoint to obtain architectural context on demand, avoiding unstructured text searches across repositories that are not checked out locally.
Checks to run with your own material and workflow
What was checked and when
Answers based on the source-checked product record
ProtoMap is a context engine that scans source code to map relationships across gRPC services and Protobuf messages, serving context to developers and coding agents.
Coding agents connect to ProtoMap via Model Context Protocol, allowing them to query caller hierarchies, dependencies, and field dataflows directly instead of grepping code.
ProtoMap can be used on a single machine, deployed on-premises using Docker, or hosted in an organization-wide cloud environment.
Users can query the blast radius of service and field changes, list all callers of an RPC method, identify unused RPCs, and determine service ownership.
Local self-hosted installation requires Git, Docker with Compose version 2.23.1 or higher, at least 4 GB of RAM, and 20 GB of available disk space.
