@leoustc/service-llm 0.2.2 → 0.2.5

package/README.md

@@ -15,6 +15,14 @@ Use your ChatGPT Codex subscription through OpenAI-compatible HTTP APIs or MCP.
 npm install --global @leoustc/service-llm
 ```
 
+Show the installed CLI version:
+
+```sh
+service-llm -v
+```
+
+Use `service-llm --help` to show all available commands and server options.
+
 ## Log in
 
 ```sh
@@ -144,9 +152,10 @@ Connect an MCP client to:
 http://127.0.0.1:7060/mcp
 ```
 
-The server provides one tool:
+The server provides two tools:
 
 - `ask`: accepts `prompt` and optional `model`, `tools`, and `tool_choice` arguments.
+- `list_models`: lists available models and reports the currently configured default model and reasoning level.
 
 Example JSON-RPC request:
 
@@ -168,6 +177,14 @@ curl http://127.0.0.1:7060/mcp \
 
 Specify a model with the request's `model` field. If it is omitted or unavailable, the service uses your configured default. On first login, the lowest-cost available model is selected with low reasoning effort. Responses always report the model actually used.
 
+List the models available to both Chat Completions and Responses clients:
+
+```sh
+curl http://127.0.0.1:7060/v1/models
+```
+
+When client authentication is enabled, include the same Bearer or `X-API-Key` header used for generation requests. The response includes `default_model` and `default_reasoning_effort` in addition to the OpenAI-compatible model list.
+
 ## Debug logging
 
 Enable request, response, and SSE logs with:

package/VERSION

@@ -1 +1 @@
-0.2.2
+0.2.5

package/dist/cli.js

@@ -1,9 +1,11 @@
 #!/usr/bin/env node
 import { listModels, login, setDefaultModel } from "./auth.js";
 import { startServer } from "./http.js";
+import { version } from "./version.js";
 
 function usage() {
   console.error(`Usage:
+  service-llm -v | --version
   service-llm login [--device-code]
   service-llm model [list]
   service-llm model set <model-id>
@@ -39,7 +41,9 @@ function valueAfter(args, names, fallback) {
 }
 
 const args = process.argv.slice(2);
-if (args.includes("--help") || args.includes("-h")) {
+if (args.length === 1 && ["-v", "--version"].includes(args[0])) {
+  console.log(`service-llm v${version}`);
+} else if (args.includes("--help") || args.includes("-h")) {
   usage();
 } else if (args[0] === "login") {
   await login({ deviceCode: args.includes("--device-code") });

package/dist/http.js

@@ -5,7 +5,6 @@ import { version } from "./version.js";
 import {
   contentFromMessage,
   createCompletionStream,
-  defaultModel,
   messagesFromChat,
   messagesFromResponses,
   toolCallsFromMessage,
@@ -26,6 +25,23 @@ function sendJson(response, status, body) {
   response.end(JSON.stringify(body));
 }
 
+export function modelCatalogResponse(catalog) {
+  const created = Math.floor(new Date(catalog.updated_at ?? 0).getTime() / 1000) || 0;
+  return {
+    object: "list",
+    data: catalog.models.map((model) => ({
+      id: model.id,
+      object: "model",
+      created,
+      owned_by: "openai-codex",
+      ...(model.name ? { name: model.name } : {}),
+    })),
+    default_model: catalog.default_model,
+    default_reasoning_effort: catalog.default_reasoning_effort ?? "low",
+    updated_at: catalog.updated_at,
+  };
+}
+
 async function readJson(request) {
   const chunks = [];
   let size = 0;
@@ -151,8 +167,8 @@ async function responses(request, response, body) {
   sendJson(response, 200, responseObject(id, model, await collect(streamResult)));
 }
 
-async function ask({ prompt, model = defaultModel, tools, toolChoice }) {
-  const completion = await createCompletionStream({
+export async function ask({ prompt, model, tools, toolChoice }, createCompletion = createCompletionStream) {
+  const completion = await createCompletion({
     context: messagesFromResponses(prompt),
     model,
     tools,
@@ -167,7 +183,20 @@ async function mcp(response, body) {
   if (body.method === "initialize") return sendJson(response, 200, { ...base, result: { protocolVersion: body.params?.protocolVersion ?? "2025-03-26", capabilities: { tools: {} }, serverInfo: { name: "service-llm", version } } });
   if (body.method === "notifications/initialized") return response.writeHead(202).end();
   if (body.method === "ping") return sendJson(response, 200, { ...base, result: {} });
-  if (body.method === "tools/list") return sendJson(response, 200, { ...base, result: { tools: [{ name: "ask", description: "Ask a Codex model a question and optionally allow it to return function tool calls. Tool calls are never executed.", inputSchema: { type: "object", properties: { prompt: { type: "string" }, model: { type: "string" }, tools: { type: "array", items: { type: "object" } }, tool_choice: {} }, required: ["prompt"], additionalProperties: false } }] } });
+  if (body.method === "tools/list") return sendJson(response, 200, { ...base, result: { tools: [
+    { name: "ask", description: "Ask a Codex model a question and optionally allow it to return function tool calls. Tool calls are never executed.", inputSchema: { type: "object", properties: { prompt: { type: "string" }, model: { type: "string" }, tools: { type: "array", items: { type: "object" } }, tool_choice: {} }, required: ["prompt"], additionalProperties: false } },
+    { name: "list_models", description: "List the available Codex models and report the model currently configured as the default.", inputSchema: { type: "object", properties: {}, additionalProperties: false } },
+  ] } });
+  if (body.method === "tools/call" && body.params?.name === "list_models") {
+    const catalog = modelCatalogResponse(await listModels());
+    return sendJson(response, 200, {
+      ...base,
+      result: {
+        content: [{ type: "text", text: JSON.stringify(catalog, null, 2) }],
+        structuredContent: catalog,
+      },
+    });
+  }
   if (body.method === "tools/call" && body.params?.name === "ask") {
     if (typeof body.params.arguments?.prompt !== "string") return sendJson(response, 200, { ...base, error: { code: -32602, message: "ask requires a string prompt" } });
     const answer = await ask({
@@ -209,6 +238,7 @@ export function startServer({ port = 7060, host = "127.0.0.1", apiKeys = [] } =
       });
       if (request.method === "GET" && url.pathname === "/health") return sendJson(response, 200, { status: "ok" });
       if (!isAuthorized(request, allowedKeys)) return sendJson(response, 401, { error: { message: "Invalid API key", type: "authentication_error" } });
+      if (request.method === "GET" && url.pathname === "/v1/models") return sendJson(response, 200, modelCatalogResponse(await listModels()));
       if (request.method !== "POST") return sendJson(response, 405, { error: { message: "Method not allowed" } });
       const body = await readJson(request);
       debug("request body", body);
@@ -245,6 +275,7 @@ export function startServer({ port = 7060, host = "127.0.0.1", apiKeys = [] } =
     }
     console.error("Endpoints:");
     console.error(`  Health:           ${baseUrl}/health`);
+    console.error(`  Models:           ${baseUrl}/v1/models`);
     console.error(`  Chat Completions: ${baseUrl}/v1/chat/completions`);
     console.error(`  Responses:        ${baseUrl}/v1/responses`);
     console.error(`  MCP:              ${baseUrl}/mcp`);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@leoustc/service-llm",
-  "version": "0.2.2",
+  "version": "0.2.5",
   "description": "Standalone CLI service exposing a Codex subscription through OpenAI-compatible and MCP endpoints.",
   "type": "module",
   "main": "./dist/index.js",