@artinet/fleet 0.1.1 → 0.1.7

package/README.md

@@ -9,7 +9,16 @@
 
 Deploy AI agents on any infrastructure.
 
-Fleet is a lightweight server framework for hosting agents with built-in orchestration, tool integration (MCP), and Agent2Agent communication.
+Fleet is a lightweight server framework for hosting agents with built-in orchestration, tool integration (MCP), and Agent2Agent (A2A) communication.
+
+## Features
+
+- **Multi-Framework Support** — Express and Hono adapters (with edge runtime compatibility on the way).
+- **A2A Protocol Compliant** — Full JSON-RPC 2.0 implementation for agent interactions
+- **MCP Integration** — Connect to Model Context Protocol servers for tool access
+- **Pluggable Storage** — In-memory, SQLite, or custom storage backends
+- **Custom Middleware** — Intercept and transform requests/responses
+- **Multi-Tenant Ready** — Built-in user isolation for SaaS deployments
 
 ## Installation
 
@@ -17,40 +26,38 @@ Fleet is a lightweight server framework for hosting agents with built-in orchest
 npm install @artinet/fleet openai @modelcontextprotocol/sdk @a2a-js/sdk
 ```
 
-**Requirements:** Node.js ≥ 18.9.1
+**Requirements:** Node.js >= 18.9.1
 
 ## Quick Start
 
 ### 1. Launch a Fleet
 
-Set an `OPENAI_API_KEY` in you environment variables, then start your Server.
+Set an `OPENAI_API_KEY` in your environment variables, then start your server.
 
-**Express**:
+**Express:**
 
 ```bash
 npm install express
 ```
 
 ```typescript
-import { fleet } from "@artinet/fleet/express";
+import { fleet } from '@artinet/fleet/express';
 
 fleet().launch(3000);
 ```
 
-**Hono**:
+**Hono:**
 
 ```bash
 npm install hono
 ```
 
 ```typescript
-import { fleet } from "@artinet/fleet/hono";
+import { fleet } from '@artinet/fleet/hono';
 
 fleet().launch(3000);
 ```
 
-> 🚧 **More servers coming soon** — Bun adapters and edge support are on the roadmap.
-
 That's it. You now have:
 
 - `POST /deploy` — Deploy agents
@@ -60,79 +67,96 @@ That's it. You now have:
 
 ### 2. Deploy an Agent
 
-**Prelaunch**:
+**Prelaunch (ship before launch):**
 
 ```typescript
-import { fleet } from "@artinet/fleet/express";
+import { fleet } from '@artinet/fleet/express';
 
 const myFleet = await fleet().ship([
-  {
-    config: {
-      uri: "my-agent",
-      ...
+    {
+        config: {
+            uri: 'my-agent',
+            name: 'My Agent',
+            description: 'A helpful assistant',
+            modelId: 'gpt-4o',
+            instructions: 'You are a helpful assistant.',
+            version: '1.0.0',
+            skills: [],
+            capabilities: {},
+            defaultInputModes: ['text'],
+            defaultOutputModes: ['text'],
+            services: [],
+        },
     },
-  },
 ]);
 
 myFleet.launch(3000);
 ```
 
-**Post Launch, Ship**:
+**Post Launch (ship to a running server):**
+
+Use the ship command, it uses zod to verify agent configurations before sending them to fleet.
 
 ```typescript
-import { ship } from "@artinet/fleet";
+import { ship } from '@artinet/fleet';
 
-await ship("http://localhost:3000", {
+await ship('http://localhost:3000', {
     config: {
-      uri: "my-agent",
-      ...
+        uri: 'my-agent',
+        name: 'My Agent',
+        description: 'A helpful assistant',
+        modelId: 'gpt-4o',
+        instructions: 'You are a helpful assistant.',
+        version: '1.0.0',
+        skills: [],
+        capabilities: {},
+        defaultInputModes: ['text'],
+        defaultOutputModes: ['text'],
+        services: [],
     },
 });
 ```
 
-**Curl**:
+**With MCP Tools:**
 
-```bash
-curl -X POST http://localhost:3000/deploy \
-  -H "Content-Type: application/json" \
-  -d '{
-    "config": {
-      "name": "My Agent",
-      "uri": "my-agent",
-      "description": "A helpful assistant",
-      "modelId": "gpt-4",
-      "instructions": "You are a helpful assistant.",
-      "version": "1.0.0",
-      "skills": [],
-      "capabilities" : {},
-      "defaultInputModes": ["text"],
-      "defaultOutputModes": ["text"],
-      "services": [{
-        "type": "mcp",
-        "uri": "everything-server-1",
-        "info": {
-          "implementation": {
-            "version": "0.0.1",
-            "name": "everything"
-          }
-        },
-        "arguments": {
-          "command": "npx",
-          "args": [
-            "-y",
-            "@modelcontextprotocol/server-everything@2025.11.25"
-          ]
-        }
-      }]
-    }
-  }'
-```
+```typescript
+import { ship } from '@artinet/fleet';
 
-> 🚧 Coming Soon: Support for Remote MCP Servers.
+await ship('http://localhost:3000', {
+    config: {
+        uri: 'tool-agent',
+        name: 'Tool Agent',
+        description: 'An agent with access to tools',
+        modelId: 'gpt-4o',
+        instructions: 'You are a helpful assistant with tool access.',
+        version: '1.0.0',
+        skills: [],
+        capabilities: {},
+        defaultInputModes: ['text'],
+        defaultOutputModes: ['text'],
+        services: [
+            {
+                type: 'mcp',
+                uri: 'everything-server',
+                info: {
+                    implementation: {
+                        version: '0.0.1',
+                        name: 'everything',
+                    },
+                },
+                arguments: {
+                    command: 'npx',
+                    args: ['-y', '@modelcontextprotocol/server-everything@2025.11.25'],
+                },
+            },
+        ],
+    },
+});
+```
 
 ### 3. Talk to Your Agent
 
-via curl:
+**Via curl:**
 
 ```bash
 curl -X POST http://localhost:3000/agentId/my-agent \
@@ -152,177 +176,34 @@ curl -X POST http://localhost:3000/agentId/my-agent \
   }'
 ```
 
-or via the [sdk](https://github.com/the-artinet-project/artinet-sdk):
+**Via the SDK:**
 
 ```typescript
-import { createMessenger } from "@artinet/sdk";
+import { createMessenger } from '@artinet/sdk';
 
 const messenger = createMessenger({
-  baseUrl: "http://localhost:3000/agentId/my-agent",
+    baseUrl: 'http://localhost:3000/agentId/my-agent',
 });
 
 // Send a message
-const response = await messenger.sendMessage("Hello!");
-
+const response = await messenger.sendMessage('Hello!');
 console.log(response);
 
 // Or stream the response
-for await (const update of messenger.sendMessageStream("Tell me a story")) {
-  console.log(update);
+for await (const update of messenger.sendMessageStream('Tell me a story')) {
+    console.log(update);
 }
 ```
 
-### SQLite Storage
-
-Set up a SQLite Database with [drizzle](https://www.npmjs.com/package/drizzle-orm):
-
-```bash
-npm install drizzle-orm better-sqlite3
-```
-
-```typescript
-import { SQLiteStore, AgentsTable } from "@artinet/fleet/sqlite";
-import { fleet } from "@artinet/fleet/hono";
-/*Use any drizzle compatible Database*/
-import Database from "better-sqlite3";
-import { drizzle } from "drizzle-orm/better-sqlite3";
-
-const sqlite = new Database("fleet.db");
-const db = drizzle<AgentsTable>(sqlite);
-
-fleet({
-  storage: new SQLiteStore(db),
-}).launch(3000);
-```
-
-### Logging
-
-Setup a custom logger via the [@artinet/sdk](https://www.npmjs.com/package/@artinet/sdk):
-
-```bash
-npm install @artinet/sdk pino pino-pretty
-```
-
-```typescript
-import { configure } from "@artinet/sdk";
-import { configurePino } from "@artinet/sdk/pino";
-import pino from "pino";
-
-configure({
-  logger: configurePino(
-    pino({
-      level: "info",
-      transport: {
-        target: "pino-pretty",
-        options: { colorize: true },
-      },
-    })
-  ),
-});
-```
-
-### Middleware
-
-Intercept and transform agent requests and responses by adding `Middleware`:
-
-```typescript
-import { fleet } from "@artinet/fleet/express";
-import { Middleware } from "@artinet/fleet";
-
-fleet({
-  middleware: new Middleware()
-    .request(async ({ request, context }) => {
-      // Inspect or transform incoming requests
-      console.log("Incoming request:", request);
-      return request;
-    })
-    .response(
-      async ({ response, context }) => {
-        // Inspect or transform outgoing responses
-        console.log("Outgoing response:", response);
-        return response;
-      },
-      // Use a trigger function to determine if the middleware should fire (defaults to `true` for every request/response)
-      ({ response, context }) => {
-        return true;
-      }
-    ),
-}).launch(3000);
-```
-
-The middleware chain is composable & sequential; add multiple `request` or `response` handlers as needed. Each handler receives the current request/response and context, and must return the (optionally modified) value.
-
-## [Docker Configuration](https://github.com/the-artinet-project/artinet/blob/main/fleet/dockerfile)
-
-Build the docker image:
-
-```bash
-docker build -t artinet-fleet .
-```
-
-Copy the example and fill in your values:
+## Documentation
 
-```bash
-cp .env.example .env
-# Edit .env with your API keys
-```
-
-Run:
-
-```bash
-docker run --env-file .env -v fleet-data:/data -p 3000:3000 -e PORT=3000 artinet-fleet
-```
-
-<!-- ## Custom Handlers
-
-```typescript
-import { fleet, InMemoryStore } from "@artinet/fleet/express";
-
-const storage = new InMemoryStore();
-
-const app = fleet(
-  {
-    storage,
-
-    // Custom agent retrieval
-    get: async (request, context) => {
-      const agent = await storage.get(context.agentId);
-      // ... return response
-    },
-
-    // Custom deployment logic
-    set: async (request, context) => {
-      await storage.set(request.config.uri, request.config);
-      return { agentId: request.config.uri, success: true };
-    },
-  },
-  {
-    // Optional auth middleware
-    auth: async (req, res, next) => {
-      if (validateToken(req.headers.authorization)) {
-        next();
-      } else {
-        res.status(401).json({ error: "Unauthorized" });
-      }
-    },
-  }
-);
-
-app.listen(3000);
-``` -->
-
-## Settings
-
-| Option                 | Type         | Default                        | Description                                                                                                                     |
-| ---------------------- | ------------ | ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------- |
-| `storage`              | `IDataStore` | `InMemoryStore`, `SQLiteStore` | Agent storage backend (storage adapters coming soon)                                                                            |
-| `basePath`             | `string`     | `"/"`                          | Base path for all routes                                                                                                        |
-| `agentPath`            | `string`     | `"/agentId"`                   | Agent interaction path                                                                                                          |
-| `deploymentPath`       | `string`     | `"/deploy"`                    | Deployment endpoint                                                                                                             |
-| `testPath`             | `string`     | `"/test"`                      | Test endpoint                                                                                                                   |
-| `inferenceProviderUrl` | `string`     | `undefined`                    | An OpenAI API compatible endpoint                                                                                               |
-| `load`                 | `function`   | `loadAgent`                    | Returns an A2A Protocol compliant agent wrapped in the [`@artinet/sdk`](<(https://github.com/the-artinet-project/artinet-sdk)>) |
-| `middleware`           | `Middleware` | `undefined`                    | Request/response interceptors for the agent route                                                                               |
+| Document                           | Description                        |
+| ---------------------------------- | ---------------------------------- |
+| [Settings](./docs/settings.md)     | Complete settings reference        |
+| [Storage](./docs/storage.md)       | Storage backends and configuration |
+| [Middleware](./docs/middleware.md) | Request/response interception      |
+| [API Reference](./docs/api.md)     | Endpoints and JSON-RPC methods     |
+| [Deployment](./docs/deployment.md) | Docker and production deployment   |
 
 ## API Reference
 
@@ -331,7 +212,7 @@ app.listen(3000);
 | Method | Path                                       | Description          |
 | ------ | ------------------------------------------ | -------------------- |
 | POST   | `/deploy`                                  | Deploy a new agent   |
-| POST   | `/test`                                    | Test a new agent     |
+| POST   | `/test`                                    | Test an agent        |
 | GET    | `/agentId/:id/.well-known/agent-card.json` | Get agent card       |
 | POST   | `/agentId/:id`                             | JSON-RPC interaction |
 
@@ -343,25 +224,23 @@ app.listen(3000);
 | `message/stream` | Send a message, stream events |
 | `task/get`       | Get task status               |
 | `task/cancel`    | Cancel a running task         |
-| `resubscribe`    | stream events                 |
+| `resubscribe`    | Stream events                 |
 
 ## Architecture
 
 ```
-
 @artinet/fleet
-├── /express # Express adapter
-├── /hono # Hono adapter
-└── /bun # Coming soon
+├── /express     # Express adapter
+├── /hono        # Hono adapter
+└── /sqlite      # SQLite storage adapter
 
 Depends on:
-├── @artinet/armada # Core business logic
-├── @artinet/sdk # A2A protocol client/server
-├── orc8 # Agent/Tool orchestration
-├── agent-def # Standardized Agent Definitions
-├── openai # OpenAI API Client
-└── @mcp # @modelcontextprotocol/sdk
-
+├── @artinet/armada   # Core business logic
+├── @artinet/sdk      # A2A protocol client/server
+├── orc8              # Agent/Tool orchestration
+├── agent-def         # Standardized Agent Definitions
+├── openai            # OpenAI API Client
+└── @mcp              # @modelcontextprotocol/sdk
 ```
 
 ## Testing
@@ -376,4 +255,4 @@ Apache-2.0
 
 ## Contributing
 
-Contributions welcome! Please open an issue or PR on [GitHub](https://github.com/the-artinet-project/fleet).
+Contributions welcome! Please open an issue or PR on [GitHub](https://github.com/the-artinet-project/artinet).

package/dist/default.d.ts

@@ -692,3 +692,4 @@ export declare const DEFAULTS: {
     storage: InMemoryStore;
     baseUrl: string;
 };
+export declare const AGENT_FIELD_NAME = "agentId";

package/dist/default.js

@@ -15,3 +15,4 @@ export const DEFAULTS = {
     storage: new InMemoryStore(),
     baseUrl: "http://localhost:3000",
 };
+export const AGENT_FIELD_NAME = 'agentId';