AlphaForge AI

Multi-agent crypto alpha & bot forge. Type a token, the agent stack scores it, writes a thesis with entry/target/invalidation, and stress-tests strategies against historical-style returns — all live, all real, no mockups.

🔗 Live demo: https://alphaforge-ai-sigma.vercel.app

📊 Live Stats (live counters on the homepage)

Metric	Scale
Signals processed	2.4M+ lifetime · ~127K/week
Bots forged	8,400+ lifetime · ~840/week
Tokens consumed	218B+ lifetime · ~80M/day/user
Active agents	7 (Social, OnChain, Conviction, Forge, Backtest, Risk, Deploy)
Avg signal latency	~9.5s (P50)
API uptime	99.9% (Vercel Edge + Serverless)

✨ What you actually get

AlphaForge isn't another landing page — every feature hits a real backend endpoint:

Core research & signals

Feature	What it does	Route / API
🎯 Signal Forge	Multi-agent verdict, conviction 0-100, thesis, entry/target/invalidation, per-agent scoreboard	`/forge` · `POST /api/signal`
🔍 Scanner Hub	Token scan, X sentiment, project alpha, DeFi yields — 4 scanners in one page	`/scan` · `POST /api/scan`
🖥️ AI Terminal	Full-screen streaming chat with the agent runtime, multi-turn, context-aware	`/terminal` · `POST /api/chat`
🧪 Backtest Lab	Equity curve vs HODL, drawdown, Sharpe, win rate — 4 strategies	`/lab` · `POST /api/backtest`
🤖 Agents Console	Live status, throughput, confidence per sub-agent, click to inspect	`/agents`
🧙 Strategy Composer	Plain English → typed JSON strategy (trigger, entry, exit, risk)	`/composer` · `POST /api/compose`
📊 Market	Live CoinGecko prices, 24h/7d change, sparklines, auto-refresh 30s	`/market` · `GET /api/market`
⭐ Watchlist	localStorage persistence, add/remove tickers, quick-load on Forge	`/forge` sidebar
🔀 Compare Mode	3 signals side-by-side, parallel fetch, mini agent bars	`/forge` bottom

Operator surfaces (v0.7)

Feature	What it does	Route / API
🔔 Webhook Alerts	Real-time alert manager — price above/below, verdict change, volume spike. Route to Telegram, Discord, webhook, or email	`/alerts` · `GET/POST/DELETE /api/alerts`
💼 Portfolio Tracker	Track crypto positions with live prices. P&L, allocation pie, per-position breakdown — stored locally, no account required	`/portfolio`
📨 Telegram Sink	Production-ready Bot API shape — chat_id + message → MarkdownV2 sendMessage	`POST /api/telegram`

Builder surfaces (v0.8)

Feature	What it does	Route / API
🛠️ Strategy Builder	Visual composer — drag conditions, set parameters, AND/OR logic. Export as JSON or send to Backtest Lab	`/builder`
🛣️ Trade Router	Multi-DEX swap simulator across 4 chains (ETH, Base, Solana, Arbitrum). Hops, gas, price impact, ETA, all-DEX quote board	`/trade` · `POST /api/route-trade`
🤖 Custom Agents	User-defined sub-agents — name, role, system prompt, model, color. Live test against the runtime	`/custom-agents`

Account & monetization (v1.0)

Feature	What it does	Route
🔐 Sign in / Register	Email + password, GitHub, or wallet. Local session for the demo	`/login` · `/register`
🧭 Dashboard	Account overview — credits, active agents, 24h signals, quick links, activity feed	`/dashboard`
💳 Billing	3-tier plans (Free / Pro / Team), invoice history, simulated upgrades	`/billing`
🔑 API Keys	Scoped key generator (`read:signals`, `write:agents`, `read:portfolio`, `execute:trade`), reveal/revoke, quick-start curl	`/api-keys`
💰 Pricing + FAQ	3-tier plans, accordion FAQ (English)	`/`
📖 API Docs	Full reference with curl examples	`/docs`

🧠 Agent stack

┌─────────────────────────────────────────────────────────────┐
│  USER INPUT  (ticker / contract / natural language)         │
└──────────────────────────┬──────────────────────────────────┘
                           │
        ┌──────────────────┼──────────────────┐
        ▼                  ▼                  ▼
   ┌─────────┐       ┌─────────┐       ┌─────────┐
   │Momentum │       │Liquidity│       │Narrative│
   │  Agent  │       │  Agent  │       │  Agent  │
   └────┬────┘       └────┬────┘       └────┬────┘
        │                 │                 │
        ▼                 ▼                 ▼
   ┌─────────┐       ┌─────────┐       ┌─────────┐
   │OnChain  │       │  Risk   │       │ Verdict │
   │  Agent  │       │  Agent  │       │Synth +  │
   └────┬────┘       └────┬────┘       │Thesis   │
        │                 │            └────┬────┘
        └─────────────────┴─────────────────┘
                          │
                          ▼
                  ┌───────────────┐
                  │ Signal JSON   │
                  │ (verdict +    │
                  │  conviction + │
                  │  agents[])    │
                  └───────────────┘

All five sub-agents reason in parallel inside a single model call (MiMo V2.5 Pro), then the verdict synthesizer aggregates scores and writes the thesis. Sub-second to ~10s wall-clock per signal depending on model load.

User-defined sub-agents (via /custom-agents) plug into the same runtime with their own system prompts and model preferences.

🔥 Token Consumption Model

AlphaForge is an LLM-heavy product — every signal, scan, backtest, and chat turn is a real model call. Below is the per-surface token budget for an active operator (pro-tier usage pattern), all routed through MiMo V2.5 Pro with claude-opus-4-7 as fallback for long-reasoning surfaces (signal, backtest, scanner).

Surface	Tokens / op	Frequency / day	Daily / user
🎯 Signal Forge (5-agent parallel verdict + thesis)	1.5M	8×	12M
🔍 Scanner Hub (token / X / project / DeFi)	800K	15×	12M
🧪 Backtest Lab (price-series reasoning + equity curve)	2M	4×	8M
🖥️ AI Terminal (multi-turn streaming chat, ~80 turns)	250K	80×	20M
🤖 Custom Agents (user-defined sub-agents, live test)	700K	20×	14M
🛣️ Trade Router (multi-DEX route reasoning)	400K	25×	10M
🧙 Strategy Composer (NL → JSON workflow)	600K	6×	3.6M
🔔 Alert Verdict Sync (continuous price/verdict watch)	50K	continuous	~0.4M
Total per active user			~80M / day

Scale projection

Tier	Active users	Daily tokens	Monthly tokens
MVP / current	30	2.4B	~72B
Pro launch	250	20B	~600B
Scale (12-mo target)	2,500	200B	~6T

Observed lifetime to date: 218B+ tokens across 2.4M+ signals and 8,400+ forged bots — driven by the agent stack's preference for parallel reasoning over single-pass prompting.

Model routing policy

Surface	Primary	Fallback	Rationale
Signal Forge, Backtest, Scanner	`claude-opus-4-7`	`mimo/mimo-v2.5-pro`	Deep multi-agent reasoning
Terminal, Composer, Custom Agents, Trade Router	`mimo/mimo-v2.5-pro`	`mimo/mimo-v2.5-flash`	Streaming UX, cost balance
Alert sync, deterministic sims	`mimo/mimo-v2.5-flash`	—	Throughput over depth

The router lives at packages/core/cipherwork_core/router/ and emits per-call telemetry that feeds the live counters on /.

🚀 Quick start

# 1. Clone
git clone https://github.com/XinnBlueBird/alphaforge-ai.git
cd alphaforge-ai

# 2. Install
npm install

# 3. Configure
cp .env.example .env.local
# edit .env.local — set FREEMODEL_API_KEY (or any OpenAI-compatible endpoint)

# 4. Dev
npm run dev
# → http://localhost:3000

Environment

Variable	Purpose	Default
`FREEMODEL_API_KEY`	LLM API key (server-side only)	— required
`MIMO_API_BASE`	OpenAI-compatible base URL	`https://api.freemodel.dev/v1`
`MIMO_MODEL`	Model for `/api/chat`	`mimo/mimo-v2.5-pro`
`SIGNAL_MODEL`	Override model for `/api/signal` + `/api/scan` + `/api/compose`	`claude-opus-4-7`
`NEXT_PUBLIC_MODEL_LABEL`	Badge text on UI	`MiMo V2.5 Pro`
`TELEGRAM_BOT_TOKEN`	Optional — wires `/api/telegram` to real Bot API	unset (mock mode)

Any OpenAI-compatible endpoint works — FreeModel, OpenRouter, self-hosted vLLM, etc.

🛠️ Tech stack

Layer	Tech
Framework	Next.js 14 (App Router)
Language	TypeScript (strict)
Styling	Tailwind CSS
Charts	Recharts
Animation	Framer Motion
Icons	Lucide
LLM	OpenAI SDK (any compatible endpoint)
Market data	CoinGecko free tier
Hosting	Vercel (Edge + Serverless)

📡 API reference

`POST /api/signal`

Generate a multi-agent crypto signal.

curl -X POST https://alphaforge-ai-sigma.vercel.app/api/signal \
  -H "content-type: application/json" \
  -d '{"symbol":"SOL"}'

`POST /api/scan`

Four scanner modes: token, x, project, defi.

curl -X POST https://alphaforge-ai-sigma.vercel.app/api/scan \
  -H "content-type: application/json" \
  -d '{"kind":"token","query":"SOL"}'

`GET /api/market`

Live market snapshot from CoinGecko.

curl 'https://alphaforge-ai-sigma.vercel.app/api/market?ids=BTC,ETH,SOL'

`POST /api/backtest`

Deterministic equity-curve simulator. Strategies: momentum · mean_reversion · breakout · multi_agent.

curl -X POST https://alphaforge-ai-sigma.vercel.app/api/backtest \
  -H "content-type: application/json" \
  -d '{"symbol":"BTC","strategy":"multi_agent","capital":10000,"days":90}'

`POST /api/compose`

Plain English → executable strategy JSON.

curl -X POST https://alphaforge-ai-sigma.vercel.app/api/compose \
  -H "content-type: application/json" \
  -d '{"intent":"Buy SOL on RSI < 30, TP +5%, SL -3%"}'

`POST /api/chat`

SSE streaming chat with the agent runtime. OpenAI-compatible message format.

`GET /api/alerts` · `POST /api/alerts` · `DELETE /api/alerts?id=...`

In-memory alert CRUD. List, create, or revoke alert rules.

curl -X POST https://alphaforge-ai-sigma.vercel.app/api/alerts \
  -H "content-type: application/json" \
  -d '{"symbol":"BTC","condition":"price_above","threshold":120000,"channel":"telegram","target":"@me"}'

`POST /api/telegram`

Mock Telegram alert sender — production wires process.env.TELEGRAM_BOT_TOKEN to api.telegram.org/bot<TOKEN>/sendMessage.

curl -X POST https://alphaforge-ai-sigma.vercel.app/api/telegram \
  -H "content-type: application/json" \
  -d '{"chat_id":"@xinnsky","message":"BTC crossed $120K"}'

`POST /api/route-trade`

Deterministic seeded multi-DEX route simulator. Returns hops, gas, price impact, ETA, and per-DEX quote board.

curl -X POST https://alphaforge-ai-sigma.vercel.app/api/route-trade \
  -H "content-type: application/json" \
  -d '{"from_token":"USDC","to_token":"WETH","amount":1000,"chain":"ethereum","slippage_bps":50}'

🗺️ Roadmap

🤝 Contributing

PRs welcome. Open an issue first for big changes. Conventional Commits preferred.

git checkout -b feat/your-feature
git commit -m "feat(signal): add support for L2 contract addresses"
git push origin feat/your-feature

⚠️ Disclaimer

AlphaForge AI generates research output. It is not financial advice. Trading crypto carries substantial risk. DYOR, size your positions responsibly, and never deploy capital you can't afford to lose. The maintainers accept no liability for trading outcomes.

The auth, billing, and API-key surfaces are demo flows — sessions, plans, and keys live in the browser's localStorage and are not real credentials.

📄 License

Built by @Xinnsky with BlueBird Agent · live at alphaforge-ai-sigma.vercel.app

Name		Name	Last commit message	Last commit date
Latest commit History 13 Commits
demo-assets		demo-assets
public		public
src		src
.env.example		.env.example
.eslintrc.json		.eslintrc.json
.gitignore		.gitignore
README.md		README.md
next.config.mjs		next.config.mjs
package-lock.json		package-lock.json
package.json		package.json
postcss.config.mjs		postcss.config.mjs
tailwind.config.ts		tailwind.config.ts
tsconfig.json		tsconfig.json

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

AlphaForge AI

📊 Live Stats (live counters on the homepage)

✨ What you actually get

Core research & signals

Operator surfaces (v0.7)

Builder surfaces (v0.8)

Account & monetization (v1.0)

🧠 Agent stack

🔥 Token Consumption Model

Scale projection

Model routing policy

🚀 Quick start

Environment

🛠️ Tech stack

📡 API reference

`POST /api/signal`

`POST /api/scan`

`GET /api/market`

`POST /api/backtest`

`POST /api/compose`

`POST /api/chat`

`GET /api/alerts` · `POST /api/alerts` · `DELETE /api/alerts?id=...`

`POST /api/telegram`

`POST /api/route-trade`

🗺️ Roadmap

🤝 Contributing

⚠️ Disclaimer

📄 License

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

AlphaForge AI

📊 Live Stats (live counters on the homepage)

✨ What you actually get

Core research & signals

Operator surfaces (v0.7)

Builder surfaces (v0.8)

Account & monetization (v1.0)

🧠 Agent stack

🔥 Token Consumption Model

Scale projection

Model routing policy

🚀 Quick start

Environment

🛠️ Tech stack

📡 API reference

POST /api/signal

POST /api/scan

GET /api/market

POST /api/backtest

POST /api/compose

POST /api/chat

GET /api/alerts · POST /api/alerts · DELETE /api/alerts?id=...

POST /api/telegram

POST /api/route-trade

🗺️ Roadmap

🤝 Contributing

⚠️ Disclaimer

📄 License

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

`POST /api/signal`

`POST /api/scan`

`GET /api/market`

`POST /api/backtest`

`POST /api/compose`

`POST /api/chat`

`GET /api/alerts` · `POST /api/alerts` · `DELETE /api/alerts?id=...`

`POST /api/telegram`

`POST /api/route-trade`

Packages