The FastEdge MCP server is the executor layer for AI-assisted FastEdge development. It exposes Model Context Protocol tools that any MCP-compatible client can call directly — Cursor, VS Code, Claude Desktop, Codex, and Claude Code (via the gcore-fastedge plugin).
What it does:
- Containerised build toolchains for JS/TS, Rust, and AssemblyScript (no local toolchain needed) —
build-wasm,upload-binary - Schema-aware Gcore API access —
gcore_api,describe_api,workflows_list,batch_execute - Bundled FastEdge reference docs kept current via the plugin's docs pipeline —
fastedge-docs - Project scaffolding from FastEdge SDK templates —
scaffold-fastedge-project,list-fastedge-templates
Higher-level workflow orchestration — deciding when to build, what to deploy, how to wire CDN rules — lives in the gcore-fastedge plugin skills. This server runs the steps.
If you're on Claude Code or Codex, use the plugin instead. It wraps this server with the intelligence layer (blueprint-driven scaffolding, deploy/manage/live-test/debug skills) and is the recommended path:
- Claude Code →
gcore-fastedgeplugin - Codex →
gcore-fastedge-codexplugin (same repo)
The rest of this README covers running the MCP server standalone for clients without plugin support (Cursor, Claude Desktop, VS Code MCP, etc.).
- Wire the Docker image into your client's MCP config — see STANDALONE-SETUP.md.
- Invoke the
createFastEdgeAppprompt to scaffold your first project. (The exact form depends on your client —/createFastEdgeApp,/mcp__fastedge-assistant__createFastEdgeApp, or a UI prompt picker.)
To get comprehensive information about FastEdge development:
Show me FastEdge development patterns and best practices
Get FastEdge context for building edge applications
To scaffold a new FastEdge project using the interactive prompt:
/createFastEdgeApp
For a complete deployment workflow using the prompt:
/deployFastEdgeApp
This automated workflow will:
- Build the current code into a WASM binary using
build-wasm - Upload the binary to FastEdge using
upload-binary - Use
describe_apito inspect thefastedge-appsendpoints, then callgcore_apito create or update the application (POST/fastedge/v1/appsor PUT/fastedge/v1/apps/{id}) with the new binary ID - Optionally generate Magic Comments for tracking deployment info
If you prefer to execute steps individually:
Build WASM binary:
Build my current file into a WASM binary
Upload binary:
Upload the generated WASM binary to FastEdge
Deploy application:
Create a FastEdge application using binary ID 12345 with name "my-app"
To manually add "Magic Comments" use:
/insertMagicComments
To manually deploy environment variables from dotenv files use:
/setEnvironmentVariables
For more information on how Environment Variables work in dotenv files read here
- scaffold-fastedge-project - Create a new FastEdge project from templates.
- list-fastedge-templates - List all available FastEdge templates with descriptions, languages, and application types.
- build-wasm - Build a FastEdge WASM binary from source code within the workspace.
- deployment-comments - Generate "Magic Comments" for tracking deployment information within code files.
- fastedge-docs - Search and read FastEdge SDK reference docs, platform guides, error codes, and testing documentation. Three modes: topics (list all), search (keyword match), read (full document).
- upload-binary - Upload a WASM binary to the FastEdge API for deployment.
- gcore_api - Execute any Gcore API call. Use
describe_apifirst to understand available endpoints. - describe_api - Get endpoint documentation and TypeScript type definitions for a Gcore API resource group (e.g.
fastedge-apps,cdn-resources). Covers FastEdge, CDN, DNS, WAAP, and Storage. - workflows_list - Discover pre-built multi-step API workflows (e.g.
create-app,update-app-binary). Returns templates forbatch_execute. - batch_execute - Execute multiple sequential API calls in one invocation with
$name.pathreference resolution between steps. Max 5 calls per batch (configurable viaBATCH_MAX_CALLS); total runtime capped at 3 minutes.
MCP resources are read-only guidance documents that agents can read on demand.
- fastedge-scaffolding-guide (
fastedge://guides/scaffolding) — Decision tree and recipes for scaffolding FastEdge apps across new repos, mixed-language monorepos, and existing projects. - fastedge-template-guide (
fastedge://guides/templates) — Template selection guide coveringhttp-base,http-react,http-react-hono, andcdn-base.
Up-to-date SDK, platform, and example documentation is served through the fastedge-docs tool (above) rather than as a static resource.
- createFastEdgeApp — Intelligent project creation that detects new vs existing repos, mixed-language scenarios, and routes to the right scaffold action
- deployFastEdgeApp — Build, upload, and deploy via
build-wasm+upload-binary+describe_api/gcore_api - setEnvironmentVariables — Collect environment variables, secrets, and response headers from
.envfiles and apply them to a deployed app viagcore_api - insertMagicComments — Generate Magic Comments for deployment tracking
- explainFastEdgeTemplate — Walk through the available templates and recommend one for the user's use case
Make sure to set the following environment variables:
GCORE_API_KEY(required) - Your Gcore API key for authentication (legacyFASTEDGE_API_KEYalso accepted).GCORE_API_BASE(optional) - Runtime override for the Gcore API base URL. Defaults tohttps://api.gcore.com(baked at build time). In-house devs can set this tohttps://api.preprod.worldto test against preprod endpoints using prod schemas.BATCH_MAX_CALLS(optional) - Override the default max calls perbatch_execute(default: 5).
See DEVELOPMENT.md for the full env var table and the preprod build recipe.
The MCP server includes the following templates:
- http-base - Basic HTTP request/response handling for JavaScript/TypeScript
- http-react - React application template for static site hosting
- http-react-hono - React application with Hono framework for enhanced routing
- cdn-base - CDN application template for proxy/traffic modification (AssemblyScript)
Each template supports multiple programming languages including JavaScript, TypeScript, and AssemblyScript.
The system supports "Magic Comments" that track deployment information directly in your source code:
/* FastEdge Deployment Magic Comments
* appName: "my-application"
* appId: "12345"
* appUrl: "https://my-app.fastedge.app"
* outputFile: "/wasm/output.wasm"
* buildDirectory: "./dist"
*/These comments are automatically used by the build and deployment tools to maintain consistency across deployments.
When developing FastEdge applications:
- Use the Service Workers API pattern for HTTP event handling
- Keep code lightweight and minimize dependencies for optimal edge performance
- Use async/await for asynchronous operations
- Implement proper error handling with appropriate HTTP status codes
- Test locally before deploying to the edge
- Use environment variables and secrets for configuration management
- Follow FastEdge examples for proven patterns and best practices
- JavaScript/TypeScript: Full support with FastEdge SDK and modern framework integration
- AssemblyScript: WebAssembly-first development for CDN applications
- Rust: High-performance edge applications (examples and documentation provided)
This MCP server is consumed by, and integrates with:
gcore-fastedgeplugin for Claude Code — primary consumer; ships a bundled.mcp.jsonthat launches this server and delegates build/deploy/manage execution to itgcore-fastedge-codexplugin — Codex variant of the same plugin, also launches this server- Gcore API — direct REST access via
gcore_api(FastEdge, CDN, DNS, WAAP, Storage) - FastEdge SDKs (
@gcoredev/fastedge-sdk-js,fastedgeRust crate,@gcoredev/proxy-wasm-sdk-as) — bundled in the container's build toolchain - FastEdge Examples Repository — source of truth for the bundled reference docs and scaffold blueprints
- VS Code FastEdge Launcher Extension — generates the
.vscode/mcp.jsonconfig that points at this server
- FastEdge Documentation: https://g-core.github.io/FastEdge-sdk-js/
- FastEdge Examples: https://github.com/G-Core/FastEdge-examples
- Model Context Protocol: https://modelcontextprotocol.io/