Two MCP tools that replace hundreds. Give an AI agent your OpenAPI spec and a request handler — it discovers and calls your entire API by writing JavaScript in a sandboxed runtime.
Instead of defining individual MCP tools for every API endpoint (list-pods, create-product, get-logs, ...), CodeMode exposes just two tools:
search— the agent writes JS to filter your OpenAPI spec and discover endpointsexecute— the agent writes JS to call your API via an injected client
This is the same pattern Cloudflare uses to expose 2,500+ API endpoints through just two MCP tools, reducing context window usage by 99.9%.
Requires mise for tooling (Node.js, pnpm, Task):
git clone https://github.com/cnap-tech/codemode.git
cd codemode
mise install # installs Node 24, pnpm 10, Task
task install # installs dependencies
task example # runs the Petstore demoFetches the real Petstore OpenAPI spec from the web, then runs search + execute against a local Hono mock — no API keys needed.
pnpm add @robinbraemer/codemode
# Install a sandbox runtime (at least one):
pnpm add isolated-vm # V8 isolates — recommended for production on Node
pnpm add quickjs-emscripten # WASM QuickJS — fallback for Bun / CF Workers / browserIf both are installed, the auto-selector (createExecutor) picks isolated-vm on Node and quickjs-emscripten on Bun (where isolated-vm cannot dlopen because Bun's JavaScriptCore engine does not export the V8 symbols isolated-vm requires).
import { CodeMode } from '@robinbraemer/codemode';
import { Hono } from 'hono';
const app = new Hono();
app.get('/v1/clusters', (c) => c.json([{ id: '1', name: 'prod' }]));
app.post('/v1/clusters', async (c) => {
const body = await c.req.json();
return c.json({ id: '2', ...body }, 201);
});
const codemode = new CodeMode({
spec: myOpenAPISpec, // OpenAPI 3.x spec, or async getter
request: app.request.bind(app), // in-process, no network hop
});
// The agent searches the spec to discover endpoints...
const search = await codemode.callTool('search', {
code: `async () => {
const results = [];
for (const [path, methods] of Object.entries(spec.paths)) {
for (const [method, op] of Object.entries(methods)) {
if (op.tags?.some(t => t.toLowerCase() === 'clusters')) {
results.push({ method: method.toUpperCase(), path, summary: op.summary });
}
}
}
return results;
}`
});
// ...then executes API calls
const result = await codemode.callTool('execute', {
code: `async () => {
const res = await api.request({ method: "GET", path: "/v1/clusters" });
return res.body;
}`
});import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
import { CodeMode } from '@robinbraemer/codemode';
import { registerTools } from '@robinbraemer/codemode/mcp';
const codemode = new CodeMode({
spec: () => fetchOpenAPISpec(),
request: app.request.bind(app),
});
const server = new McpServer({ name: 'my-api', version: '1.0.0' });
registerTools(codemode, server);
const transport = new StdioServerTransport();
await server.connect(transport);AI Agent
│ writes JavaScript code
▼
CodeMode MCP Server
│
├─ search(code) → runs JS with preprocessed OpenAPI spec
│ → all $refs resolved inline, only essential fields kept
│ → agent discovers endpoints, schemas, parameters
│
└─ execute(code) → runs JS with injected request client
→ api.request() calls your handler in-process
→ no network hop, auth handled automatically
All code runs in an isolated V8 sandbox. The sandbox has zero I/O by default — no require, no process, no fetch, no filesystem. The only way to interact with the outside world is through the injected globals (spec for search, {namespace}.request() for execute).
Each tool call gets a fresh sandbox with no state carried over between calls.
| Option | Type | Default | Description |
|---|---|---|---|
spec |
OpenAPISpec | () => OpenAPISpec | Promise<OpenAPISpec> |
required | OpenAPI 3.x spec or async getter |
request |
(input, init?) => Response |
required | Fetch-compatible handler (app.request.bind(app) for Hono) |
namespace |
string |
"api" |
Client name in sandbox (api.request(...)). Must be a valid JS identifier, not a reserved name. |
baseUrl |
string |
"http://localhost" |
Base URL for relative paths |
sandbox |
SandboxOptions |
see below | Sandbox resource limits |
executor |
Executor |
IsolatedVMExecutor |
Custom sandbox executor |
maxResponseTokens |
number |
25000 |
Token limit for response truncation (0 to disable) |
maxRequests |
number |
50 |
Max requests per execute() call |
maxResponseBytes |
number |
10485760 |
Max response body size in bytes (10MB) |
allowedHeaders |
string[] |
undefined |
Header whitelist. When unset, a blocklist strips Authorization, Cookie, Host, X-Forwarded-*, Proxy-*. |
maxRefDepth |
number |
50 |
Max $ref resolution depth |
| Option | Type | Default | Description |
|---|---|---|---|
memoryMB |
number |
64 |
V8 isolate memory limit |
timeoutMs |
number |
30000 |
CPU timeout in ms (caps pure compute) |
wallTimeMs |
number |
60000 |
Wall-clock timeout in ms (caps total elapsed time including async I/O) |
Returns MCP-compatible tool definitions for search and execute.
Route a tool call. Returns { content: [{ type: "text", text }], isError? }.
Run search code directly (shorthand for callTool('search', { code })).
Run execute code directly (shorthand for callTool('execute', { code })).
Override default tool names. Useful when running multiple CodeMode instances.
Clean up sandbox resources.
The spec global is the preprocessed OpenAPI spec with all $ref pointers resolved inline:
// Find endpoints by tag
async () => {
const results = [];
for (const [path, methods] of Object.entries(spec.paths)) {
for (const [method, op] of Object.entries(methods)) {
if (op.tags?.some(t => t.toLowerCase() === 'clusters')) {
results.push({ method: method.toUpperCase(), path, summary: op.summary });
}
}
}
return results;
}
// Get endpoint with requestBody schema (refs are already resolved)
async () => {
const op = spec.paths['/v1/products']?.post;
return { summary: op?.summary, requestBody: op?.requestBody };
}
// Spec metadata
async () => ({
title: spec.info.title,
version: spec.info.version,
endpoints: Object.keys(spec.paths).length,
})The {namespace}.request() function makes API calls through the host handler:
// GET with query params
async () => {
const res = await api.request({
method: "GET",
path: "/v1/clusters",
query: { limit: 10 },
});
return res.body;
}
// POST with body
async () => {
const res = await api.request({
method: "POST",
path: "/v1/products",
body: { name: "Redis", chart: "bitnami/redis" },
});
return { status: res.status, body: res.body };
}
// Chain calls
async () => {
const list = await api.request({ method: "GET", path: "/v1/clusters" });
const details = await Promise.all(
list.body.map(c =>
api.request({ method: "GET", path: `/v1/clusters/${c.id}` })
)
);
return details.map(d => d.body);
}Request options:
| Field | Type | Description |
|---|---|---|
method |
string |
HTTP method ("GET", "POST", etc.) |
path |
string |
API path ("/v1/clusters") |
query |
Record<string, string | number | boolean> |
Query parameters (optional) |
body |
unknown |
Request body, auto-serialized as JSON (optional) |
headers |
Record<string, string> |
Additional headers (optional) |
Response: { status: number, headers: Record<string, string>, body: unknown }
CodeMode automatically preprocesses your OpenAPI spec before passing it to the search sandbox:
$refresolution — all$refpointers are resolved inline (circular refs become{ $circular: ref })- Field extraction — only essential fields kept per operation:
summary,description,tags,operationId,parameters,requestBody,responses - Metadata preserved —
info,servers, andcomponents.schemasare kept alongside processed paths
You can also use the preprocessing utilities directly:
import { resolveRefs, processSpec, extractTags } from '@robinbraemer/codemode';
const processed = processSpec(rawSpec);
const tags = extractTags(rawSpec);CodeMode ships two executor backends. IsolatedVMExecutor is the recommended production backend on Node. QuickJSExecutor is a compatibility fallback for environments where isolated-vm cannot load (Bun, Cloudflare Workers, browser).
Use createExecutor() for automatic selection, or pass an executor instance explicitly:
import { CodeMode, createExecutor, IsolatedVMExecutor, QuickJSExecutor } from '@robinbraemer/codemode';
// Automatic — picks isolated-vm on Node, quickjs-emscripten on Bun
const codemode = new CodeMode({
spec,
request: handler,
executor: await createExecutor({ memoryMB: 128, timeoutMs: 60_000 }),
});
// Or explicit
const codemode = new CodeMode({
spec,
request: handler,
executor: new IsolatedVMExecutor({
memoryMB: 128,
timeoutMs: 60_000, // CPU time limit
wallTimeMs: 120_000, // total elapsed time limit
}),
});| Executor | Package | Performance | Portability | Production-ready |
|---|---|---|---|---|
IsolatedVMExecutor |
isolated-vm |
Native V8 speed | Node.js | ✅ |
QuickJSExecutor |
quickjs-emscripten |
Slower (interpreted WASM) | Node, Bun, CF Workers, browser |
- Not a production backend. Exists so the package loads on runtimes where
isolated-vmcannot dlopen. Production callers on Node should useIsolatedVMExecutor. - Sandboxed code must avoid sequential
awaiton host functions. UsePromise.all([fn1(), fn2()])for parallel calls instead. Chained sequentialawaits currently crash with an upstreamquickjs-emscripten@0.32.0release-asyncify regression (justjake/quickjs-emscripten#258) — reproduces identically on Node and Bun. - Return-value semantics differ from
isolated-vm. Host ↔ guest values cross via aJSON.stringifyenvelope.Date,Map,Set,BigIntare converted to strings/objects, not preserved as instances.isolated-vmuses structured clone and preserves them. Stick to plain JSON-safe shapes in sandboxed code that targets both backends. - CPU timeout is wall-clock-based.
isolated-vmuses true CPU time; QuickJS uses elapsed time. Async host calls that take wall time count against the CPU budget under QuickJS.
Implement the Executor interface to use your own sandbox:
import { CodeMode, type Executor, type ExecuteResult } from '@robinbraemer/codemode';
class MyExecutor implements Executor {
async execute(code: string, globals: Record<string, unknown>): Promise<ExecuteResult> {
// `code` is an async arrow function as a string: "async () => { ... }"
// `globals` contains named values to inject:
// - plain data (objects, arrays, primitives) → read-only values
// - functions → callable host functions
// - objects with function values → namespace with callable methods
return { result: ..., logs: [] };
}
dispose() { /* clean up */ }
}
const codemode = new CodeMode({
spec,
request: handler,
executor: new MyExecutor(),
});| Approach | Context Tokens |
|---|---|
| Individual MCP tools (15-50+ tools) | ~15,000-50,000+ |
| Full OpenAPI spec in context | ~1,000,000+ |
| CodeMode (2 tools) | ~1,000 |
MIT