Intelligence Terminal pulls satellite fire detection, flight tracking, radiation monitoring, satellite constellation tracking, economic indicators, live market prices, conflict data, sanctions lists, and social sentiment from 27 open-source intelligence feeds in parallel, every 15 minutes, and renders everything on a single self-contained dashboard.
Hook it up to an LLM and it becomes a **two-way intelligence assistant**: pushing multi-tier alerts to Telegram and Discord when something meaningful changes, responding to commands like `/brief` and `/sweep` from your phone, and generating trade ideas grounded in real cross-domain data.
Most of the world's real-time intelligence — satellite imagery, radiation levels, conflict events, economic indicators, flight tracking, maritime activity — is publicly available. It's just scattered across dozens of government APIs, research institutions, and open data feeds that nobody has time to check individually.
Intelligence Terminal brings it all into one place. Not behind a paywall, not locked in an enterprise platform, not requiring a security clearance. Just open data, aggregated and cross-correlated on your own machine, updated every 15 minutes.
It was built for anyone who wants to understand what's actually happening in the world right now — researchers, journalists, traders, OSINT analysts, or just curious people who believe access to information shouldn't depend on your budget.
> **If `npm run dev` fails silently** (exits with no output), run Node directly instead:
> ```bash
> node --trace-warnings server.mjs
> ```
> This bypasses npm's script runner, which can swallow errors on some systems (particularly PowerShell on Windows). You can also run `node diag.mjs` to diagnose the exact issue — it checks your Node version, tests each module import individually, and verifies port availability. See [Troubleshooting](#troubleshooting) for more.
The server starts at `http://localhost:3117` and immediately begins its first intelligence sweep. Browser auto-open is disabled by default; open the URL yourself or explicitly set `AUTO_OPEN_BROWSER=true` for a supported desktop environment. The initial sweep queries all 27 sources in parallel and typically takes 30–60 seconds — the dashboard will appear empty until the sweep completes and pushes the first data update. After that, it auto-refreshes every 15 minutes via SSE (Server-Sent Events). No manual page refresh is needed.
Dashboard at `http://localhost:3117`. Sweep data persists in `./runs/` via volume mount. The container disables browser auto-open by default, exposes `/api/health` and `/api/metrics`, and is suitable for Dockge/Pangolin.
For Pangolin or another reverse proxy, forward HTTP traffic to `intelligence-terminal:3117` (or the `PORT` you set). Missing API keys do not crash sweeps; affected sources are reported as degraded in `/api/health`.
`POST /api/action` and `POST /api/sweep` can trigger operational actions such as manual sweeps. The dashboard has a **SET TOKEN** control that stores your `SWEEP_TOKEN` in browser local storage and sends it as the `x-crucix-token` header; do not put action tokens in URLs.
| Private local machine | `NODE_ENV=development`, optional `SWEEP_TOKEN`, optional `TERMINAL_ACTIONS_ENABLED=true`. Localhost can run actions without a token for development. |
| Private LAN / Dockge | Set a strong `SWEEP_TOKEN`, keep `TERMINAL_ACTIONS_ENABLED=true`, expose only to trusted clients. |
| Pangolin-authenticated reverse proxy | Set a strong `SWEEP_TOKEN`, keep Pangolin auth in front, use the dashboard **SET TOKEN** flow once per browser. |
| Public internet | Do not expose Terminal Actions directly. If exposure is unavoidable, require `SWEEP_TOKEN`, keep proxy authentication enabled, lower `TERMINAL_ACTION_RATE_LIMIT_MAX`, and monitor server audit logs. |
Action endpoints reject cross-origin POST origins, apply a small in-memory per-IP rate limit, and write sanitized audit lines without logging the token.
When data remains stale past `STALE_DATA_MAX_AGE_MINUTES`, the server sends an operator alert through configured Telegram/Discord channels after failed or degraded sweep attempts. `STALE_ALERT_COOLDOWN_MINUTES` prevents repeated stale alerts from spamming every refresh interval. Set `DASHBOARD_URL` to the Pangolin/public URL you want included in those alerts.
Crucix stores longitudinal memory in `runs/intelligence.db` when the current Node.js build exposes `node:sqlite`. If SQLite is unavailable, the file is created as a harmless placeholder and `/api/health` reports the memory store as unavailable instead of failing the sweep.
The memory layer persists:
| Table | Purpose |
| --- | --- |
| `runs` | Sweep timestamps, source health counts, and delta direction summaries. |
| `entities` | Stable entity IDs for recurring countries, regions, and locations. |
| `events` | Stable event IDs for conflict, OSINT, urgent news, and new delta signals across sweeps. |
| `predictions` | Trade/intelligence hypotheses with evidence, confidence, horizon, outcome state, and latest grading. |
Query endpoints:
```text
GET /api/memory/search?q=iran&limit=25
GET /api/memory/predictions?state=open&limit=25
```
Memory endpoints use the same operator authorization gate as Terminal Actions. The dashboard Terminal Actions panel includes a `Memory` action for a quick operator-facing view of recent events and prediction states.
Retention, backup, and privacy expectations:
- Treat `runs/intelligence.db` as operator data. It can contain source excerpts, headlines, generated hypotheses, and URLs from your configured feeds.
- Back up `runs/` with the rest of your Dockge volume if you want longitudinal learning to survive container replacement.
- Delete `runs/intelligence.db` to reset SQLite memory; the next sweep recreates the schema.
- Do not commit `runs/` or `.env`. API credentials stay in `.env`; memory stores derived observations, not secrets.
- If you expose the dashboard through a reverse proxy, protect Terminal Actions and memory queries behind your normal authentication boundary.
The dashboard receives live sweep updates from `GET /events` using Server-Sent Events. The server sends `retry: 10000` reconnect guidance and lightweight heartbeat comments every `SSE_HEARTBEAT_INTERVAL_MS` milliseconds so reverse proxies do not close an otherwise idle stream between 15-minute sweeps.
Recommended proxy settings:
| Proxy | Setting |
| --- | --- |
| Pangolin / Traefik-style frontends | Keep response streaming enabled and set idle timeouts above `SSE_HEARTBEAT_INTERVAL_MS`. |
| Nginx | Disable proxy buffering for `/events`, keep `proxy_read_timeout` above the heartbeat interval, and preserve `Connection: keep-alive`. |
| Cloudflare-style proxies | Keep the heartbeat below common idle cutoffs; the default 25s is intentionally conservative. |
If you raise the heartbeat interval, keep it shorter than the lowest idle timeout in the proxy chain.
`/api/metrics` includes network health grouped by host and source/provider. Source modules should use `safeFetch(url, { source: 'SourceName' })`; when omitted, the shared helper infers a stable provider bucket from the URL host instead of grouping normal source traffic under `unknown`. Raw fetch exceptions are documented in [Source Fetch Instrumentation](docs/source-fetch-instrumentation.md).
Intelligence Terminal can track operator hypotheses across sweeps with a runtime scenario file at `runs/scenarios.json`. On first run, the server creates three disabled starter examples:
"invalidation": "WTI normalizes and urgent regional signals fade."
}
]
}
```
Malformed scenario config degrades safely: sweeps continue and the dashboard shows the watchlist as a config issue. Scenario state is persisted in `runs/scenario-state.json`; delete that file to reset state transitions without deleting definitions.
Scenario states are `dormant`, `watching`, `building`, and `confirmed`. The dashboard shows active scenario state, confidence, score, and recent trigger time. Briefings include a `Scenario Watchlist` section when one or more scenarios change state.
Gitea Actions publishes the same image automatically when the repository secret `REGISTRY_TOKEN` is set with package read/write permissions. The workflow tags images as `latest`, the commit SHA, and a UTC `YYYYMMDD` release tag.
- **3D WebGL globe** (Globe.gl) with atmosphere glow, star field, and smooth rotation — plus a classic flat map toggle
- **9 marker types** across both views: fire detections, air traffic, radiation sites, maritime chokepoints, SDR receivers, OSINT events, health alerts, geolocated news, conflict events
- **Animated 3D flight corridor arcs** between air traffic hotspots and global hubs
- **Region filters** (World, Americas, Europe, Middle East, Asia Pacific, Africa) — rotates the globe or zooms the flat map
- **Live market data** — indexes, crypto, energy, commodities via Yahoo Finance (no API key needed)
The `VISUALS FULL` / `VISUALS LITE` button in the top bar only changes rendering behavior - it does **not** remove data sources or reduce sweep coverage.
5. Evaluates breaking news alerts — multi-tier (FLASH / PRIORITY / ROUTINE) with semantic dedup. Sends to Telegram and/or Discord if configured. Works with LLM evaluation or falls back to rule-based alerting when LLM is unavailable.
Normal text messages in the configured private chat are treated as AI questions, so commands are optional. Answers include a compact snapshot of the latest sweep, recent ideas, evidence links, degraded sources, and a bounded conversation history. Snapshot fields are treated as untrusted evidence rather than instructions. Conversation history remains in memory only and is cleared on restart or with `/reset`.
This requires `TELEGRAM_BOT_TOKEN`, `TELEGRAM_CHAT_ID`, and a configured LLM in `.env`. The bot ignores every other chat ID and polls every 5 seconds by default. For group chats, BotFather privacy settings may prevent the bot from receiving normal text; private chat is recommended. Local reasoning models can take several minutes, so the bot refreshes Telegram's typing indicator while waiting.
When `TELEGRAM_AGENT_ENABLED=true`, the chat can perform up to `TELEGRAM_AGENT_MAX_STEPS` structured tool calls before answering. The allowlist includes system status, the approved Security Manager profile, latest brief, sweep delta, markets, source health, local evidence search, bounded public web/news search through GDELT, memory, predictions, scenarios, and generated ideas. The agent has no generic shell, filesystem, environment, secret-access, or arbitrary network tool.
Read-only tools run automatically. `trigger_sweep`, `mute_alerts`, and `unmute_alerts` return an expiring confirmation request with Telegram **Confirm** and **Cancel** buttons. Confirmation is bound to the configured chat and cannot be reused. `/trace` exposes only tool names, duration, status, and a short operational rationale; private chain-of-thought is neither requested nor stored.
With `TELEGRAM_AGENT_PROACTIVE_ENABLED=true`, material sweep changes trigger a separate bounded analysis. The Security Manager can cross-check the approved profile, evidence, source health, scenarios, memory, predictions, and bounded GDELT web/news results before deciding whether to notify. A cooldown limits repeat notifications. Without a completed profile, fixed alert rules remain the fallback; with a profile, successful agent decisions respect its alert level and quiet hours. FLASH alerts can override quiet hours.
Set a unique `SECURITY_PROFILE_ENCRYPTION_KEY` of at least 32 characters and keep the `runs` volume persistent. On the first Telegram startup, the bot asks for the language before asking for any personal context. It then presents a consent notice and offers full, minimal, or cancelled setup.
The optional profile is limited to a preferred name, country/region/city-level location, timezone, household composition, mobility, general travel pattern, risk priorities, critical dependencies, alert level, and quiet hours. Do not enter an exact address, identity documents, passwords, API keys, account details, detailed diagnoses, booking data, or private contact details. `/skip` skips an input. The profile is written only after explicit review confirmation, stored as AES-256-GCM ciphertext in `runs/security-profile.enc`, and exposed only to the configured LLM through the read-only `get_security_profile` agent tool. `/profile_delete` removes it locally after confirmation.
The Security Manager uses the profile to rank proximity, severity, time horizon, household impact, mobility constraints, and infrastructure dependencies. Its prompt requires separation of verified facts, official guidance, source reports, and inference. It is an intelligence aid, not an emergency service; urgent responses direct the operator to appropriate local authorities without claiming guaranteed safety.
The Security Manager's conversational identity is **DAVE**, a synthetic security-management construct. DAVE adapts language, formality, directness, answer length, formatting, and technical depth to the operator's current message and bounded chat history. Explicit style requests take priority. The adaptation does not copy spelling mistakes, hostility, panic, or unsupported certainty, and urgent safety instructions remain concise and unambiguous. DAVE does not claim human emotions, consciousness, a body, or access beyond the terminal's allowlisted tools.
DAVE Telegram answers are sent as plain-text chat messages. Long AI or proactive answers are split into smaller conversational messages so Telegram does not show giant report blocks or visible Markdown markers such as `**bold**`. The first proactive message should carry the immediate assessment; evidence and tool traces are sent as separate short follow-ups when present.
The language selected during Security Manager onboarding is enforced across direct chat, multi-tool answers, tool-loop finalization, dynamic presence, and proactive sweep alerts. This does not depend on the model deciding to read the profile first. Server-generated labels such as priority, evidence, and tool traces are localized as well.
`DAVE_PRESENCE_ENABLED=true` lets DAVE initiate Telegram conversations without fixed daily times. Evaluations occur at randomized intervals and can be pulled forward by material sweep changes. DAVE considers current data freshness, source integrity, recent changes, evidence, bounded web/news corroboration when useful, the encrypted Security Manager profile, time since your last interaction, and whether another message would add value. It may send a relevant warning, situation update, evidence-grounded all-clear, practical suggestion, or one natural context question. It stays silent when a message would only add noise.
Dynamic presence respects profile quiet hours, a daily message cap, a minimum gap between messages, and an idle period after your own chat activity. Evaluation timing and counters persist in `runs/dave-presence-state.json`, preventing restart spam. Scheduled presence cannot execute mutating tools or bypass confirmation. This feature is opt-in because it consumes LLM requests and sends unsolicited Telegram messages.
Intelligence Terminal also supports Discord as a full-featured bot with slash commands and rich embed alerts. It mirrors the Telegram bot's capabilities with Discord-native formatting.
| `/status` | System health, last sweep time, source status, LLM status |
| `/sweep` | Trigger a manual sweep cycle |
| `/brief` | Compact text summary of the latest intelligence |
| `/portfolio` | Portfolio status (if Alpaca connected) |
Alerts are delivered as rich embeds with color-coded sidebars: red for FLASH, yellow for PRIORITY, blue for ROUTINE. Each embed includes signal details, confidence scores, and cross-domain correlations.
**Setup requires:** `DISCORD_BOT_TOKEN`, `DISCORD_CHANNEL_ID`, and optionally `DISCORD_GUILD_ID` for instant slash command registration. See [API Keys Setup](#api-keys-setup) for details.
**Webhook fallback:** If you don't want to run a full bot, set `DISCORD_WEBHOOK_URL` instead. This enables one-way alerts (no slash commands) with zero dependencies — no `discord.js` needed.
**Optional dependency:** The full bot requires `discord.js`. Install it with `npm install discord.js`. If it's not installed, Intelligence Terminal automatically falls back to webhook-only mode.
| `REDDIT_CLIENT_ID` + `REDDIT_CLIENT_SECRET` | Reddit social sentiment | [reddit.com/prefs/apps](https://www.reddit.com/prefs/apps/) — create a script app |
Reddit is OAuth-only in this fork. If the Reddit credentials are missing or rejected, the Reddit source is reported as degraded and no unauthenticated `reddit.com/.../hot.json` fallback is used.
- GDELT web/news pulls are bounded so optional geo enrichment cannot consume the full sweep source budget.
- OFAC sanctions metadata is read through bounded XML snippets instead of downloading full sanctions XML files during every sweep.
- EPA RadNet uses shorter bounded requests and reports degraded state instead of blocking the sweep when the government endpoint is slow.
- BLS automatically retries without `BLS_API_KEY` when a configured key is rejected, while surfacing a configuration warning.
- ACLED `403` means the credentials authenticated but the account still needs ACLED terms/profile/API access completed.
- ADS-B and Reddit require credentials for their production-grade feeds; missing credentials are reported as degraded/disabled rather than silently faking live data.
Set `LLM_PROVIDER` to one of: `litellm`, `openrouter`, `openai-compatible`, `lmstudio`, `ollama`, `anthropic`, `openai`, `gemini`, `codex`, `minimax`, `mistral`, or `grok`.
| `DISCORD_CHANNEL_ID` | Right-click channel in Discord (Developer Mode on) → Copy Channel ID |
| `DISCORD_GUILD_ID` | *(Optional)* Right-click server → Copy Server ID. Enables instant slash command registration (otherwise takes up to 1 hour for global commands) |
| `DISCORD_WEBHOOK_URL` | *(Optional)* Channel Settings → Integrations → Webhooks → New Webhook → Copy URL. Use this for alert-only mode without a bot |
**Discord bot setup:**
1. Go to [Discord Developer Portal](https://discord.com/developers/applications) and create a new application
2. Go to **Bot** → click **Reset Token** → copy the token to `DISCORD_BOT_TOKEN`
3. Under **Privileged Gateway Intents**, enable **Message Content Intent**
5. Copy the generated URL and open it in your browser to invite the bot to your server
6. Install the dependency: `npm install discord.js`
Alerts work with or without an LLM on both Telegram and Discord. With an LLM configured, signal evaluation is richer and more context-aware. Without one, a deterministic rule engine evaluates signals based on severity, cross-domain correlation, and signal counts.
Intelligence Terminal still works with zero API keys. 18+ sources require no authentication at all. Sources that need keys return structured errors and the rest of the sweep continues normally.
- **Minimal dependencies** — Express is the only runtime dependency. `discord.js` is optional (for Discord bot). LLM providers use raw `fetch()`, no SDKs.
Delta engine thresholds (how sensitive the system is to changes between sweeps) can be customized in `crucix.config.mjs` under the `delta.thresholds` section. The defaults are tuned to filter out noise while catching meaningful moves.
This is normal — the first sweep takes 30–60 seconds to query all 27 sources. The dashboard will populate automatically once the sweep completes. Check the terminal for sweep progress logs.
Expected behavior. Sources that require API keys will return structured errors if the key isn't set. The rest of the sweep continues normally. Check the Source Integrity section in the dashboard (or the server logs) to see which sources failed and why. The 3 most impactful free keys to add are `FRED_API_KEY`, `FIRMS_MAP_KEY`, and `EIA_API_KEY`.
OpenSky can also return `HTTP 429` when its public hotspots are queried too aggressively. Intelligence Terminal does not try to evade that limit. Instead, it surfaces the throttle/error in source health and preserves the most recent non-empty air traffic snapshot from `runs/` so the dashboard flight layer does not suddenly go blank on a throttled sweep.
Make sure `TELEGRAM_BOT_TOKEN`, `TELEGRAM_CHAT_ID`, and the LLM settings are present in `.env`. The bot only responds to the configured chat ID. Check `/status` for LLM and AI-chat state, then try `/ask What changed?`. You should see Telegram polling startup lines in the server logs. If command registration fails, verify the bot token without posting it publicly. In group chats, use `/ask` or adjust BotFather privacy settings because normal text may not be delivered to the bot.
1. Make sure `DISCORD_BOT_TOKEN` and `DISCORD_CHANNEL_ID` are set in `.env`
2. Verify `discord.js` is installed: `npm ls discord.js`. If missing, run `npm install discord.js`
3. If slash commands don't appear, set `DISCORD_GUILD_ID` — without it, global commands can take up to 1 hour to propagate. Guild-specific commands register instantly
4. Confirm the bot was invited with `bot` + `applications.commands` scopes and has `Send Messages` + `Embed Links` permissions in the target channel
5. Check server logs for `[Discord] Bot logged in as ...` on startup. If you see `[Discord] discord.js not installed`, install it and restart
6.**Webhook-only fallback:** If you just want alerts without slash commands, set `DISCORD_WEBHOOK_URL` instead of the bot token. No `discord.js` needed.
To update them: run the dashboard, wait for a sweep to complete, then use your browser's DevTools (`F12` → `Ctrl+Shift+P` → "Capture full size screenshot") or a tool like [LICEcap](https://www.cockos.com/licecap/) for GIFs.
Found a bug? Want to add a 28th source? PRs welcome. Each source is a standalone module in `apis/sources/` - just export a `briefing()` function that returns structured data and add it to the orchestrator in `apis/briefing.mjs`.
Intelligence Terminal is an AGPL-3.0-only Crucix fork focused on Docker-first home-server operation, source health transparency, Gitea Registry delivery, and operator-owned deployments. Upstream project credit remains with the original Crucix project.
