@tesseron/server

Node SDK for Tesseron. Same Zod-style action builder as @tesseron/web, but for any Node process — a CLI, daemon, worker, Express or Fastify service. Expose typed actions to MCP-compatible AI agents without opening a browser tab.

Install

npm install @tesseron/server

You also need the @tesseron/mcp gateway running locally — it's bundled inside the Claude Code plugin.

Quick start

import { tesseron } from '@tesseron/server';
import { z } from 'zod';

tesseron.app({ id: 'deployer', name: 'Deployer CLI' });

tesseron
  .action('deployService')
  .describe('Deploy a named service to the given environment.')
  .input(z.object({
    service: z.string(),
    env: z.enum(['staging', 'production']),
  }))
  .annotate({ destructive: true, requiresConfirmation: true })
  .handler(async ({ service, env }, ctx) => {
    const ok = await ctx.confirm({ question: `Deploy ${service} to ${env}?` });
    if (!ok) return { deployed: false };
    ctx.progress({ message: 'Building...', percent: 10 });
    await build(service);
    ctx.progress({ message: 'Uploading...', percent: 60 });
    await upload(service, env);
    return { deployed: true, url: urlFor(service, env) };
  });

await tesseron.connect();

What you get

• Same API as @tesseron/web — typed action builder, subscribable resources, full handler context (ctx.confirm, ctx.elicit, ctx.sample, ctx.progress, ctx.log).
• Works in any Node process — CLI tool, background daemon, HTTP server (Express, Fastify, plain http). Same action can be invoked via MCP and via your existing REST endpoints.
• Built-in WebSocket transport — NodeWebSocketTransport from ws, no bundler configuration needed.
• Typed errors — SamplingNotAvailableError, ElicitationNotAvailableError, TimeoutError, CancelledError, TransportClosedError.

Dual-channel pattern

The same action store can be exposed via MCP and a regular HTTP API:

import express from 'express';
import { tesseron } from '@tesseron/server';

const app = express();
const store = new TodoStore();

tesseron
  .action('addTodo')
  .input(z.object({ text: z.string() }))
  .handler(({ text }) => store.add(text));

app.post('/todos', (req, res) => res.json(store.add(req.body.text)));

app.listen(3000);
await tesseron.connect();

Both channels mutate the same state. See examples/express-prompts for a working version (prompt library with sampling and elicitation).

Docs

| | | |---|---| | Main repo | https://github.com/BrainBlend-AI/tesseron | | SDK reference | https://brainblend-ai.github.io/tesseron/sdk/typescript/server/ | | Protocol spec | https://brainblend-ai.github.io/tesseron/protocol/ | | Examples | https://github.com/BrainBlend-AI/tesseron/tree/main/examples |

License

Reference implementation — Business Source License 1.1 (source-available). Each release auto-converts to Apache-2.0 four years after publication.