DOCUMENTATION

Ghostboard Docs

Everything you need to set up a workspace, connect sources, and turn crypto narrative into an operator command layer.

Overview

Ghostboard is a private AI intelligence layer built for crypto teams, launch teams, and community operators. It monitors public narrative across X, Telegram, websites, and other ecosystem signals, then turns that stream into structured intelligence: summaries, alerts, content suggestions, reply drafts, launch messaging, and risk signals.

Instead of being a generic chatbot, Ghostboard acts like an always-on war room for a project. It ingests public conversation, clusters themes, detects sentiment shifts, and turns raw discussion into actionable outputs — all powered by decentralized inference through c0mpute.

privatecrypto-nativereal-timeoperator-focused

Quickstart

Spin up your first workspace in four steps. A workspace maps to a single project — its token, handles, competitors, and communities.

1 · Create a workspace

Name the workspace after your project and pick a primary token or handle to anchor tracking.

# create a workspace
ghost workspace create "Genesis Launch" --token $GHOST

2 · Add tracking targets

Track X keywords, your project handle mentions, competitor handles, and Telegram channels.

ghost track add --x "@ghostboard, #ghost, $GHOST" \
  --competitors "@rival_a, @rival_b" \
  --telegram "t.me/ghost_main"

3 · Set summary cadence

Choose how often community summaries and digests are generated: every 1h, 6h, or 24h.

ghost summary set --window 6h --digest founder

4 · Open the war room

Launch the dashboard to see the narrative feed, sentiment, radar, and alerts populate in real time.

Sources & connectors

Ghostboard ingests from the channels that actually move crypto narrative. Each source is normalized into a common signal format before processing.

  • X — keyword, handle, and cashtag mentions; reply and quote context.
  • Telegram — channel summaries and exported group discussion windows.
  • Websites & docs — project sites, launchpad pages, and configured watchlists.
  • RSS / news — ecosystem news and announcement feeds.
  • Optional wallet signals — wallet-linked chatter for advanced workspaces.

Core modules

Six modules turn raw discussion into something you can act on.

5.1 · Narrative Feed

Live feed of tracked mentions, themes, and events. Clusters similar posts into narratives instead of raw isolated mentions, and highlights spikes, recurring claims, and sudden changes in discussion volume.

5.2 · Community Summaries

Summarize Telegram groups and discussion windows every 1h / 6h / 24h. Surfaces recurring questions, complaints, support issues, and bullish / bearish themes as digests for admins and founders.

5.3 · Sentiment + Topic Intelligence

Score positive, negative, neutral, and uncertain sentiment. Break conversation into themes — launch, tokenomics, partnership, roadmap, bug report, scam concern, competitor comparison — and track how sentiment changes over time.

5.4 · Content Copilot

Generate tweet drafts, thread ideas, announcement posts, reply suggestions, and Telegram updates from live context, in multiple tones: founder, analyst, community, technical, or launch-mode.

5.5 · Alert Engine

Notify the team when a tracked keyword, wallet, competitor, or narrative crosses a threshold — sudden FUD spike, new whale mention, repeated scam accusation, or unexpected traction.

5.6 · Ask Ghostboard

Chat with your project's live narrative data over the tracked workspace.

Architecture

A high-level view of Ghostboard as five layers:

  • Ingestion — X, Telegram, websites, docs, RSS/news, launchpad pages, watchlists, optional wallet signals.
  • Processing — cleaning, deduplication, entity extraction, topic clustering, sentiment scoring, anomaly detection, event labeling.
  • Inference (c0mpute) — summarization, classification, answer generation, drafting, ranking, and agent tasks.
  • Storage & retrieval — project workspaces, historical narrative database, alerts, embeddings, search indexes.
  • Application — dashboard, AI copilot, alert center, team digest, and admin tools.

Alert engine

Alerts fire when a signal crosses a threshold you define. Each alert carries a severity, source, and a suggested next action.

# fire when FUD volume spikes on a tracked topic
ghost alert add \
  --when "sentiment.fud" --rises 200% --within 30m \
  --notify telegram,email --severity critical

Ask Ghostboard

Ask natural-language questions over your live workspace data. Example prompts:

  • "What did our community care about most in the last 24 hours?"
  • "What are the top objections to our launch?"
  • "Which competitor is getting the most positive attention this week?"

Open the Ask Ghostboard demo →

MVP scope

The first version stays narrow and proves one core promise: help a crypto team understand and respond to narrative faster.

Included

  • Workspace creation for one project.
  • X keyword, handle, and competitor tracking; Telegram export / channel summaries.
  • Daily and intraday narrative summary.
  • Basic sentiment and topic clustering.
  • Content copilot for tweet drafts and reply suggestions.
  • Simple alert engine for narrative spikes and repeated negative mentions.
  • Ask Ghostboard chat over tracked workspace data.

Deferred

  • Deep wallet intelligence and on-chain attribution.
  • Full Discord ingestion.
  • Cross-workspace benchmarking.
  • Autonomous posting / autonomous reply agents.
  • Enterprise compliance and custom deployment layers.