ETH/USDT Trading Bot Development
Budget: €30 – €250 EUR
Deliver a single ETH bot that trades ETH/USDT (spot) on Binance Testnet using my strategy: track running peak → buy on ATR-sized dip → sell only in profit after hover time.
Include a minimal Flask web dashboard + JSON API, and deploy 24/7 on Fly.io or Render with persistent storage. Code must be clean and structured so I can extend it later to 9 more bots.
---
Strategy (must match exactly)
Timeframe: 1-minute candles.
ATR: Lookback 14, multiplier k = 1.0 (UI-adjustable 1.0–1.1).
Gap clamp: Enforce configurable gap_min/gap_max (USD) for ETH to prevent unrealistic thresholds.
Entry: Maintain running peak; when (peak − current_low) ≥ ATR_gap (clamped) → market buy (respect precision/filters).
Exit (profit-only): Sell only if:
1. Hover time (minutes) satisfied and
2. price ≥ target where target covers fees + slippage + min profit.
Risk params (UI-editable): atr_lookback, atr_k, gap_min, gap_max, hover_minutes, max_open, cooldown_sec, min_profit_bps, taker_fee_bps, slip_bps.
---
Controls & Compounding
Per-bot capital (pool) and entry size (slot) — editable, persisted.
Start / Pause / Resume / Stop-after-sales (finish open positions, then stop).
Auto-compounding:
Combined intervals 30m, 1h, 1d, 1w, 1M (all active together).
30m-only mode as a separate toggle.
Compounding = move realized PnL → pool_usdt, reset realized PnL; persisted across restarts.
---
Dashboard & API (ETH only in Phase 1)
Flask single page (simple auth) showing:
Start/Pause/Resume/Stop-after-sales
Capital & Entry editors
ATR/fees/slippage/cooldown editors
Compounding toggles
Open positions count, realized PnL, link to logs
JSON API (ETH only):
GET /api/status
POST /api/start|pause|resume|stop_after_sales { "tag": "eth" }
POST /api/save_cap { "tag": "eth", "capital": 800, "entry": 40 }
POST /api/apply_intervals { "tag": "eth", "intervals": ["30m","1h","1d","1w","1M"] }
POST /api/save_atr (fields listed in Strategy section)
---
Reliability & Safety
Binance rate-limit respect; exponential backoff on HTTP-429; safe retries on transient errors.
Precision & min-size checks via exchange filters; never send orders that round to zero.
Optional auto-pause if request rate risks throttling.
Log all actions (without leaking secrets).
---
Deployment (24/7)
Python 3.12 + Flask + gunicorn container (Dockerfile).
Deploy to Fly.io (preferred) or Render:
Mount persistent volume at /data; symlink configs/ state/ control/ logs/.
Secrets (API keys, creds) via platform secret manager; no hard-coded keys.
Healthcheck and simple HTTP auth for dashboard.
---
Deliverables
1. Repo/ZIP with source, Dockerfile, entrypoint.sh, minimal UI, and code comments.
2. Running Testnet demo (Fly/Render) I can log into.
3. README covering local run, env vars, secrets, deploy, logs, update/rollback.
4. Postman collection (or curl examples) for the API.
5. Short video/screenshots: start/pause/resume/stop-after-sales, edit capital/entry & ATR, toggle compounding, and 1 buy + 1 sell on Testnet.
---
Milestones / Acceptance
M1 – Local/Testnet: UI + API work; edits persist; at least 1 buy & 1 sell on Testnet.
M2 – 24/7 Deploy: Live on Fly/Render with /data persistence; secrets set; restart/redeploy keeps state.
M3 – Handover: Final docs, code, and demo materials delivered.
---
Tech & Stack
Python 3.12, Flask, ccxt, pandas, gunicorn, Docker.
Optional: basic front-end (HTML/JS) for the single dashboard page.
Include a minimal Flask web dashboard + JSON API, and deploy 24/7 on Fly.io or Render with persistent storage. Code must be clean and structured so I can extend it later to 9 more bots.
---
Strategy (must match exactly)
Timeframe: 1-minute candles.
ATR: Lookback 14, multiplier k = 1.0 (UI-adjustable 1.0–1.1).
Gap clamp: Enforce configurable gap_min/gap_max (USD) for ETH to prevent unrealistic thresholds.
Entry: Maintain running peak; when (peak − current_low) ≥ ATR_gap (clamped) → market buy (respect precision/filters).
Exit (profit-only): Sell only if:
1. Hover time (minutes) satisfied and
2. price ≥ target where target covers fees + slippage + min profit.
Risk params (UI-editable): atr_lookback, atr_k, gap_min, gap_max, hover_minutes, max_open, cooldown_sec, min_profit_bps, taker_fee_bps, slip_bps.
---
Controls & Compounding
Per-bot capital (pool) and entry size (slot) — editable, persisted.
Start / Pause / Resume / Stop-after-sales (finish open positions, then stop).
Auto-compounding:
Combined intervals 30m, 1h, 1d, 1w, 1M (all active together).
30m-only mode as a separate toggle.
Compounding = move realized PnL → pool_usdt, reset realized PnL; persisted across restarts.
---
Dashboard & API (ETH only in Phase 1)
Flask single page (simple auth) showing:
Start/Pause/Resume/Stop-after-sales
Capital & Entry editors
ATR/fees/slippage/cooldown editors
Compounding toggles
Open positions count, realized PnL, link to logs
JSON API (ETH only):
GET /api/status
POST /api/start|pause|resume|stop_after_sales { "tag": "eth" }
POST /api/save_cap { "tag": "eth", "capital": 800, "entry": 40 }
POST /api/apply_intervals { "tag": "eth", "intervals": ["30m","1h","1d","1w","1M"] }
POST /api/save_atr (fields listed in Strategy section)
---
Reliability & Safety
Binance rate-limit respect; exponential backoff on HTTP-429; safe retries on transient errors.
Precision & min-size checks via exchange filters; never send orders that round to zero.
Optional auto-pause if request rate risks throttling.
Log all actions (without leaking secrets).
---
Deployment (24/7)
Python 3.12 + Flask + gunicorn container (Dockerfile).
Deploy to Fly.io (preferred) or Render:
Mount persistent volume at /data; symlink configs/ state/ control/ logs/.
Secrets (API keys, creds) via platform secret manager; no hard-coded keys.
Healthcheck and simple HTTP auth for dashboard.
---
Deliverables
1. Repo/ZIP with source, Dockerfile, entrypoint.sh, minimal UI, and code comments.
2. Running Testnet demo (Fly/Render) I can log into.
3. README covering local run, env vars, secrets, deploy, logs, update/rollback.
4. Postman collection (or curl examples) for the API.
5. Short video/screenshots: start/pause/resume/stop-after-sales, edit capital/entry & ATR, toggle compounding, and 1 buy + 1 sell on Testnet.
---
Milestones / Acceptance
M1 – Local/Testnet: UI + API work; edits persist; at least 1 buy & 1 sell on Testnet.
M2 – 24/7 Deploy: Live on Fly/Render with /data persistence; secrets set; restart/redeploy keeps state.
M3 – Handover: Final docs, code, and demo materials delivered.
---
Tech & Stack
Python 3.12, Flask, ccxt, pandas, gunicorn, Docker.
Optional: basic front-end (HTML/JS) for the single dashboard page.