# ClawPipe

> The intelligent AI pipeline. Cut LLM costs 30-50% with one line of code.

## What is ClawPipe?

ClawPipe is an SDK-local LLM optimization pipeline that sits between your application and LLM providers. It reduces costs 30-50% through intelligent routing, caching, context packing, and deterministic transforms -- all without changing your application logic.

ClawPipe runs locally in your process. No proxy servers, no extra hops, no added latency. Install the SDK, replace your LLM calls with `pipe.prompt()`, and start saving immediately.

## Key Features

- **Agent Booster**: Deterministic transforms that resolve prompts without calling an LLM. Math, JSON, dates, unit conversions, and UUIDs handled instantly at zero cost.
- **Context Packing**: Compress context windows by removing redundancy, deduplicating content, and stripping boilerplate. Saves 20-60% on token counts.
- **Semantic Cache**: Hash-based and embedding-based prompt deduplication. Similar prompts return cached results in milliseconds. TTL + LRU eviction.
- **Self-Learning Router**: Cost/quality/latency-aware model selection that improves with every call. Routes simple tasks to cheap models, complex ones to powerful ones.
- **Multi-Provider Gateway**: One API for OpenAI, Anthropic, DeepSeek, Mistral, Groq, and local models. Automatic failover with circuit breaker protection.
- **Swarm Orchestration**: Fan out prompts to N models in parallel. Strategies: fastest, majority vote, highest quality, or longest response.
- **RAG Pipeline**: Retrieval-augmented generation with pluggable vector stores. Automatically prepends relevant documents as context before the LLM call.
- **Offline Fallback**: Auto-detects local LLM servers (llamafile, Ollama, LM Studio). Zero-cost routing when running locally.
- **Voice Pipeline**: Speech-to-text input and text-to-speech output. Integrates with Voicebox and Whisper.cpp for free local voice processing.
- **Pipeline Tracing**: Instrument every stage with timing data. Export traces in Perfetto format for visual analysis.

## How It Works

1. Install the SDK in your language of choice.
2. Replace your LLM provider calls with `pipe.prompt()`.
3. The pipeline automatically optimizes every request: boosting, packing, caching, routing, and learning.

## Installation

### JavaScript / TypeScript
```
npm install clawpipe-ai
```

### Python
```
pip install clawpipe-ai
```

### Go
```
go get github.com/finsavvyai/clawpipe-go
```

### Rust
```
cargo add clawpipe-ai
```

### Java
```
<dependency>
  <groupId>ai.clawpipe</groupId>
  <artifactId>clawpipe-sdk</artifactId>
</dependency>
```

## Quick Start (JavaScript)

```javascript
import { ClawPipe } from 'clawpipe-ai';

const pipe = new ClawPipe({ apiKey: 'your-key' });

const result = await pipe.prompt('Explain quantum computing in one sentence');
console.log(result.text);
console.log(result.savings); // { tokens_saved: 1200, cost_saved: '$0.003', cached: false, boosted: false }
```

## Quick Start (Python)

```python
from clawpipe import ClawPipe

pipe = ClawPipe(api_key="your-key")

result = pipe.prompt("Explain quantum computing in one sentence")
print(result.text)
print(result.savings)
```

## Pipeline Stages

```
Request -> Booster -> RAG -> Pack -> Cache -> Route -> Swarm -> Call -> Learn
```

Each stage is optional and configurable. Disable any stage you do not need.

## Pricing

- **Free**: $0/month, 1,000 calls/day, all pipeline stages, 1 project
- **Pro**: $49/month, 100,000 calls/day, unlimited projects, analytics dashboard
- **Team**: $149/month, 1,000,000 calls/day, team management, SLA guarantee
- **Enterprise**: Custom pricing, unlimited calls, SSO, audit logs, dedicated infrastructure

## Supported Providers

OpenAI, Anthropic, DeepSeek, Mistral, Groq, Together AI, Fireworks AI, Ollama, llamafile, LM Studio, and any OpenAI-compatible endpoint.

## Links

- Website: https://clawpipe.ai
- Install: npm install clawpipe-ai / pip install clawpipe-ai
- Docs: https://docs.clawpipe.ai
- Dashboard: https://app.clawpipe.ai
- API: https://api.clawpipe.ai
- GitHub: https://github.com/finsavvyai/clawpipe
- npm: https://www.npmjs.com/package/clawpipe-ai
- PyPI: https://pypi.org/project/clawpipe-ai
- Status: https://status.clawpipe.ai
- Support: support@clawpipe.ai