ThinChain
The catalytic converter between raw broker data and your AI trading agent.
Battle-tested. Sanitization logic extracted from a production options trading app. Handles every bad tick Tradier has thrown at us in production.
☁️ Moving to production?
The open-source server runs locally with your own API keys. For hosted infrastructure with multi-broker failover, SLA guarantees, and webhook alerts — join the managed cloud waitlist.
The Problem
Tradier's MCP dumps 500-row JSON chains. One bad print can corrupt an agent's reasoning. ThinChain sanitizes, compresses, and circuit-breaks that data before it reaches your model.
Installation
git clone https://github.com/snipmcp/thinchain.git
cd thinchain
pip install -e ".[dev]"
cp .env.example .env
Or with Docker:
docker-compose up --build
Configuration
TRADIER_API_KEY=your_key_here
TRADIER_BASE_URL=https://api.tradier.com/v1
POLYGON_API_KEY=your_key_here
POLYGON_BASE_URL=https://api.polygon.io
DEFAULT_BROKER=tradier
CACHE_TTL_SECONDS=30
LOG_LEVEL=INFO
Usage
Three example prompts to send to Claude (or any MCP-compatible agent):
Use get_compressed_chain to get an iron condor setup on SPY expiring 2026-06-19Check if SPY options data is trustworthy before I place my tradeGet me the ATM straddle strikes for AAPL expiring 2026-06-19
Run it in two terminals
# Tab 1 — start the MCP server
python -m thinchain.server
# Tab 2 — call a tool from a Python shell or your MCP client
# Tool signatures:
# get_compressed_chain(symbol, expiration, strategy="raw",
# delta_range=None, broker="tradier")
# get_sanitized_quote(symbol, broker="tradier")
# get_circuit_status(symbol, expiration, broker="tradier")
How it works
Three layers between raw broker output and your model:
Broker API → [Sanitize] → [Compress] → [Circuit Break] → MCP Tool → AI Agent
hygiene strategy anomaly gate
rules slicing + cache
- Sanitize: Drops ghost quotes (0.0/0.0), inverted spreads, NaN/empty/dash values, clamps deltas to [-1, 1], normalizes positive theta on long options, flags illiquid strikes.
- Compress: Slices the chain to only the strikes relevant to your strategy (iron condor, straddle, etc.) — typically 95%+ token reduction.
- Circuit Break: Four-tier data quality (
clean/noisy/degraded/circuit_breaker_active). When >65% of rows are anomalous, refuses to serve stale data and falls back to last known-good cache.
Real numbers from a live SPY iron condor (2026-06-19 expiration)
RAW TRADIER PAYLOAD THINCHAIN OUTPUT
─────────────────────────────────────────────
Rows: 482 Rows: 25
Est tokens: 38,560 Est tokens: 2,000
Token savings: 95%
Anomalous rows removed: 217
Data quality: noisy → still safely served
ThinChain compresses 482 rows / 38,560 tokens down to 25 rows / 2,000 tokens — and flags 217 anomalous strikes (illiquid, ghost quotes, wide spreads) that would have polluted the agent's context.
Roadmap
- Managed cloud tier (hosted, multi-tenant, webhook alerts)
- Polygon connector hardening (live SIP feed)
- Webhook alerts for circuit breaker trips
Contributing
PRs welcome. Run pytest before submitting.
Metadata
Release files for thinchain 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| thinchain-0.1.1.tar.gz | 16.4 kB | Without provenance | Details for thinchain-0.1.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| thinchain-0.1.1-py3-none-any.whl | Python 3 | none | any | Details for thinchain-0.1.1-py3-none-any.whl |
Total release size: 31.7 kB
Release files / thinchain-0.1.1.tar.gz
| Download URL | thinchain-0.1.1.tar.gz |
|---|---|
| Size | 16.4 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
62d4614f81790d121c36837ac858449f09ea09e85863548977f69662305d1200
|
|
BLAKE2b-256 checksum How to use checksums |
523d5b1da95a2242bc7d5ea7c7d2d031d55a0cb34f2ccb4057ad817faf55f0df
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.5
|
Release files / thinchain-0.1.1-py3-none-any.whl
| Download URL | thinchain-0.1.1-py3-none-any.whl |
|---|---|
| Size | 15.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
cc462e1ef7a50c1aaa7a8b0d9c2a4c0aa780353bb1aa046354bb903da5c907e3
|
|
BLAKE2b-256 checksum How to use checksums |
5d72f6715fb9aba99d2da16ef05e611f283940cd6280ad9549e63314c5253d8a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.5
|
Security for thinchain 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.