Dronewire

A Node.js CLI tool that monitors tech products for new release notes, summarizes them with a local AI model via Ollama, and posts the summaries to Slack.

---

How it works

1. You configure one or more connectors — each one knows how to fetch release notes from a specific source
2. Dronewire runs a persistent Express server that schedules cron jobs for each connector
3. When a connector fires, it fetches new releases, filters out anything already seen, summarizes each one with Ollama, and posts to Slack
4. The CLI talks to the server to manage connectors and trigger manual checks

---

Requirements

• Node.js >= 24
• Ollama running locally (or accessible over the network)
• A Slack bot token with chat:write permission

---

Installation

npm install -g dronewire

---

Setup

1. Copy the config template

mkdir -p ~/.dronewire
cp $(npm root -g)/dronewire/dronewire.config.js ~/.dronewire/dronewire.config.js

2. Edit ~/.dronewire/dronewire.config.js

At minimum, set your Slack token and add at least one connector (see Configuration below).

3. Start the server (keep this terminal open)

dronewire server start

4. Test a connector without posting to Slack

dronewire check my-connector --dry-run

---

Configuration

The config file is a JS module at ~/.dronewire/dronewire.config.js (override with --config <path>).

export default {
  server: {
    port: 3847,
    host: 'localhost',
  },

  ollama: {
    host: 'http://localhost:11434',
    model: 'llama3.2',
    systemPrompt: 'Summarize these release notes in 2-3 concise bullet points.',
  },

  slack: {
    enabled: true,
    token: 'xoxb-your-token-here',
    centralChannel: '#releases', // all summaries go here; null to disable
  },

  connectors: [
    {
      id: 'my-wordpress',             // unique ID used in CLI commands
      type: 'wordpress',              // builtin connector type
      config: {
        url: 'https://example.com',
      },
      schedule: '0 9 * * *',          // cron schedule
      initialLookbackDate: '2025-01-01T00:00:00Z', // cutoff for first run
      slackChannel: '#my-releases',   // per-connector channel (null to skip)
      disableCentralChannel: false,   // set true to opt out of centralChannel
    },
  ],
}

Connector options

Field	Required	Description
id	Yes	Unique identifier used in CLI commands
type	Yes	wordpress (builtin) or custom
config	Yes	Connector-specific config (passed to the connector class)
schedule	Yes	Cron expression for how often to check
initialLookbackDate	No	ISO date used as the cutoff on the very first run when no state exists
slackChannel	No	Slack channel to post this connector's summaries to
disableCentralChannel	No	Set true to opt this connector out of slack.centralChannel (default: false)
disableCron	No	Set true to permanently skip scheduling, regardless of state (default: false)
systemPrompt	No	Override the global Ollama system prompt for this connector only

Slack channel routing

For each new release, summaries are posted to:

• The connector's own slackChannel (if set)
• The global slack.centralChannel (if set, and disableCentralChannel is not true)

If both are set, the summary goes to both. Set slack.disabled: true to disable all Slack output and log summaries to the console instead.

---

CLI Reference

dronewire server start

Start the Dronewire server in the foreground. Keep this running in a dedicated terminal tab.

dronewire server start
dronewire server start --config /path/to/dronewire.config.js

dronewire server status

Check whether the server is reachable.

dronewire server status

dronewire list

List all configured connectors and their current state.

dronewire list

Output columns: id, status, schedule, slack channel, last processed date

dronewire check <id>

Manually run a connector check. Runs inline by default (no server required).

dronewire check my-wordpress
dronewire check my-wordpress --dry-run          # skip Slack, log to console
dronewire check my-wordpress --via-server       # delegate to the running server

dronewire cron enable / disable

Pause or resume all cron jobs globally.

dronewire cron enable
dronewire cron disable

dronewire cron enable-connector / disable-connector <id>

Pause or resume the cron job for a specific connector. State is persisted to ~/.dronewire/dronewire.state.json.

dronewire cron enable-connector my-wordpress
dronewire cron disable-connector my-wordpress

---

Builtin Connectors

wordpress

Fetches posts from the WordPress REST API.

{
  type: 'wordpress',
  config: {
    url: 'https://example.com',   // required — base URL of the WordPress site
    perPage: 5,                   // optional — posts per request (default: 10)
    categories: [1, 2],           // optional — filter by category IDs
    tags: [3],                    // optional — filter by tag IDs
  },
}

---

Custom Connectors

Create a JS file that exports a class extending BaseConnector:

import { BaseConnector } from 'dronewire/src/connectors/BaseConnector.js';

export default class MyConnector extends BaseConnector {
  // Required — return an array of release objects
  async fetchReleases() {
    const res = await fetch('https://api.example.com/releases');
    return res.json();
  }

  // Required — return a Date for a given release
  parseReleaseDate(release) {
    return new Date(release.published_at);
  }

  // Optional — format a release for the AI prompt (defaults to JSON.stringify)
  formatForAI(release) {
    return `${release.name}\n\n${release.body}`;
  }

  // Optional — override the system prompt for this connector
  getSystemPrompt() {
    return 'Summarize this changelog entry in one sentence.';
  }
}

Register it in your config:

{
  id: 'my-service',
  type: 'custom',
  script: '/path/to/MyConnector.js',
  config: {
    // anything you want — available as this.config inside the class
  },
  schedule: '0 10 * * *',
  initialLookbackDate: '2025-01-01T00:00:00Z',
  slackChannel: '#my-service-releases',
}

---

State file

Dronewire tracks which releases have already been processed in ~/.dronewire/dronewire.state.json. You can inspect or reset it manually if needed. Per-connector cronDisabled flags written by the CLI are also stored here.

---

REST API

The server exposes a simple HTTP API that the CLI uses. You can call it directly if you want to build integrations or a web UI.

Method	Path	Description
GET	/health	Liveness check
GET	/connections	List connectors with full state
POST	/connections/:id/check	Trigger a manual check ({ dryRun: true } body optional)
GET	/crons	List cron job statuses
POST	/crons/enable	Enable global cron
POST	/crons/disable	Disable global cron
POST	/crons/:id/enable	Enable a connector's cron
POST	/crons/:id/disable	Disable a connector's cron