spirit-platform 2.2.4

Programmable cryptocurrency trading platform — write your own strategies, run on your own hardware

Spirit

Spirit is a programmable Python platform for algorithmic cryptocurrency trading — a playground for building and testing your own strategies. It runs on your own hardware, talks to exchanges directly, and you own the strategy logic end-to-end.

The framework is open source under Apache-2.0. Spirit handles the orchestration: market data, order placement, lifecycle hooks, and crash recovery. As your strategies mature, Plus and Pro open up access to our custom technical indicators and cloud storage for backtesting.

---

Quick start

On Ubuntu / Debian (and most modern Linux):

sudo apt install -y pipx
pipx ensurepath
pipx install spirit-platform

Then open a new shell and run:

python3 -m spirit.setup     # interactive setup wizard
spirit --mode paper         # start paper trading

Other install paths (venv, etc.) and troubleshooting: see INSTALL.md.

Run the wizard once (it asks your tier, instance name, and an optional Kraken API key for live trading), pick one of the bundled examples (sma_crossover or macd_demo) or drop your own file in ~/.spirit/strategies/, and Spirit starts paper-trading.

To go live, swap --mode paper for --mode live once you've sanity-checked the paper P&L.

OHLC data ownership (v2.2.4+). Spirit's local OHLC store is yours — Free instances persist candles in ~/.spirit/<instance>/spirit.db. When Spirit starts up after downtime, it fills the gap (up to 12 hours of 1-minute candles per boot) from the exchange automatically. For longer gaps, run python3 -m spirit.backfill <kraken-csv> once with a Kraken CSV export. The boot catch-up adds ~15–20 seconds for a typical multi-pair multi-interval config; configure with SPIRIT_OHLC_CATCHUP_INTERVALS (default 60).

---

Tiers and pricing

See www.tradebot.live/pricing for tier details, pricing, and what each level adds.

---

Writing a strategy

Strategies subclass BaseStrategy and implement evaluate_trade(). Everything else is optional; opt into lifecycle hooks as you need them.

from spirit.strategies.base import BaseStrategy

class MyStrategy(BaseStrategy):
    def evaluate_trade(self, pair: str, mode: str = "test", **kwargs):
        # Return {"entry": bool, "exit": bool, "details": {...}}
        return {"entry": False, "exit": False, "details": {}}

Optional hooks the orchestrator will call when configured: on_monitoring_tick, on_entry_confirmed, on_exit_completed, validate_readiness, get_data_requirements. Properties for tier-aware behaviour: uses_risk_gate, required_capabilities. See the bundled examples; both files are heavily commented teaching artifacts:

• src/spirit/strategies/examples/sma_crossover.py: minimum viable. Subclass + evaluate_trade, nothing else. Read this first.
• src/spirit/strategies/examples/macd_demo.py: full lifecycle tour. Multi-interval, monitoring-tick ATR stop, entry-confirmed state-stash, paper-by-default guard. Read this to see every hook in action.

Drop your own under ~/.spirit/strategies/ and Spirit picks it up at next startup.

For more information visit www.tradebot.live.

---

Project structure

src/spirit/
  main.py                  - entrypoint
  setup.py                 - first-run wizard
  config.py                - .env / yaml loader
  trade_signal.py          - signal dataclass
  trade_status.py          - status dataclass
  trade_types.py           - TradeRecord dataclass (used by strategies)
  exchange/                - exchange adapter protocol + Kraken impl
  pipeline/                - WebSocket event bus, freshness cache, daemon health
  storage/                 - local SQLite schema (Free tier)
  strategies/
    base.py                - BaseStrategy abstract base class
    examples/sma_crossover - minimal reference strategy
    examples/macd_demo     - full-lifecycle reference strategy
  utils/                   - data providers, OHLC buffer, paper executor, etc.

The framework runs any strategy you write against any data source you plug in. Plus and Pro plans add bundled indicators served via the gateway API — the D-Limit suite, V3 confidence scorer, and risk-gate calibrators — without changing the framework you're building against. See tradebot.live for what each tier unlocks.

---

Supported exchanges

Spirit includes a Kraken adapter by default. You can write your own by implementing the ExchangeProvider protocol in src/spirit/exchange/protocol.py — see docs/reference/EXCHANGE_PLUGIN_GUIDE.md for the full guide.

We'll be releasing more exchange adapters over time. If there's a specific exchange you'd like us to prioritise, open a GitHub issue and we'll see if we can fit it in.

---

Status

Current release: see the latest tag on PyPI or GitHub Releases for the changelog. The Apache-2.0 framework is stable; Plus and Pro infrastructure is in active development.

---

License

Apache-2.0. See LICENSE.

The framework is free to use, modify, and redistribute. Plus and Pro subscriptions cover the hosted data and indicator infrastructure; the framework code itself doesn't depend on a subscription to run.

---

Support

• Portal + key management: portal.tradebot.live
• Issue tracker: GitHub Issues on this repo
• For commercial / integration questions: [e-mail]