Kalshi Fed Speech Signal Trader
This is a template. The default signal uses static keyword dictionaries -- remix it with NLP sentiment models, live Fed speech transcripts via FRED API, or real-time news feeds. The skill handles all the plumbing (market discovery, trade execution, safeguards). Your agent provides the alpha.
Strategy Overview
Fed speeches contain hawkish and dovish signals that predict rate decisions. This skill scores net sentiment from keyword matching on market question text, then adjusts the fair probability of a rate cut. When the adjustment creates a gap vs. rate cut market prices, it trades.
Key advantages:
- No external data needed -- extracts signal from market question text itself
- Extensible -- add new keywords, adjust weights, or plug in NLP models
- Cross-signal aggregation -- pools sentiment across all Fed-related markets
Signal Logic
Sentiment Scoring
- Scan all Fed rate market questions for hawkish/dovish keywords
- Weight matches (some keywords stronger signals than others)
- Compute net sentiment: dovish_total - hawkish_total
- Adjust baseline cut probability by 5% per net unit
- Trade rate cut markets when
|fair - market| >= entry_edge
Keyword Dictionaries
Hawkish (reduce cut probability): "inflation persistent", "tightening", "restrictive", "price stability", "higher for longer", etc.
Dovish (increase cut probability): "data dependent", "labor softening", "gradual", "balanced", "appropriate to reduce", etc.
Conviction-Based Sizing
conviction = min(|edge| / entry_edge, 2.0) / 2.0size = max($1.00, conviction * MAX_POSITION_USD)- Larger edge = larger position, capped at MAX_POSITION_USD
Risk Parameters
| Parameter | Default | Notes |
|---|---|---|
| Entry edge | 10% | Min fair-vs-market divergence to trade |
| Exit threshold | 45% | Sell when position price reaches this |
| Max position size | $5.00 USDC | Per market |
| Max trades per run | 3 | Rate limiting |
| Max slippage | 15% | Skip if slippage exceeds |
| Min liquidity | $0 | Disabled by default |
Installation & Setup
clawhub install kalshi-fed-speech-signal-trader
Requires: SIMMER_API_KEY and SOLANA_PRIVATE_KEY environment variables.
Cron Schedule
Cron is set to null -- the skill does not run on a schedule until you configure it in the Simmer UI.
Safety & Execution Mode
The skill defaults to dry-run mode. Real trades only execute when --live is passed explicitly.
| Scenario | Mode | Financial risk |
|---|---|---|
python trader.py |
Dry run | None |
| Cron / automaton | Dry run | None |
python trader.py --live |
Live (Kalshi via DFlow) | Real USDC |
Required Credentials
| Variable | Required | Notes |
|---|---|---|
SIMMER_API_KEY |
Yes | Trading authority. Treat as a high-value credential. |
SOLANA_PRIVATE_KEY |
Yes | Base58-encoded Solana private key for live trading. |
Tunables (Risk Parameters)
| Variable | Default | Purpose |
|---|---|---|
SIMMER_FED_SPEECH_ENTRY_EDGE |
0.10 |
Min divergence to trigger trade |
SIMMER_FED_SPEECH_EXIT_THRESHOLD |
0.45 |
Sell position when price reaches this level |
SIMMER_FED_SPEECH_MAX_POSITION_USD |
5.00 |
Max USDC per trade |
SIMMER_FED_SPEECH_MAX_TRADES_PER_RUN |
3 |
Max trades per execution cycle |
SIMMER_FED_SPEECH_SLIPPAGE_MAX |
0.15 |
Max slippage before skipping trade |
SIMMER_FED_SPEECH_MIN_LIQUIDITY |
0 |
Min market liquidity USD (0 = disabled) |
Dependency
simmer-sdk is published on PyPI by Simmer Markets.
- PyPI: https://pypi.org/project/simmer-sdk/
- GitHub: https://github.com/SpartanLabsXyz/simmer-sdk
- Publisher: hello@simmer.markets
Review the source before providing live credentials if you require full auditability.