lightweight llm pipeline library

Generate publication-grade academic figures from a methodology brief via a 6-phase drawio + LLM pipeline.

Governed structured-output LLM pipeline: rate limit → spend caps (per-user + global circuit breaker) → cache → provider-chain failover → Zod validation → usage logging → LLM-judge, with a deterministic mock provider for CI.

YAML DSL for LLM pipeline creation. Tiny. ESM. No deps.

Orchestrates the query -> retrieve -> explain -> LLM pipeline for conversational graph queries.

create a pipeline of streams using Minipass

Transform pipeline operator into call expressions

Core HTTP Compatibility Library to bridge the gap between Core V1 & V2 packages.

AWS SDK for JavaScript Codepipeline Client for Node.js, Browser and React Native

TypeScript primitive for verifying contracts between LLM pipeline steps. Typed inputs/outputs via Zod, quality gates with retry, and failure context that makes silent agent failure impossible.

Provider-neutral LLM service interface for the DeepSeek Harness

Allow parsing of the pipeline operator

Provider-routed LLM request retry policy for the DeepSeek Harness

The default executor-less/UI-less agent spine with fallback session titles, provider-routed retry, and optional persisted goals

Three-agent LLM pipeline in your terminal: a router classifies the prompt, a worker generates, a reviewer audits — all on free Groq + Gemini tiers with local budget tracking

AI red-teaming for the AI agents & LLM apps you own: stress-test them with adversarial models to find security failures (prompt injection, tool misuse / excessive agency, data leakage, jailbreaks, denial-of-wallet), auto-patch (blue team), retest, and get

Hardware accelerated language model chats on browsers

Chain functions, generators, Node streams, and Web streams into a pipeline with backpressure support.

pi-ai-backed DeepSeek adapter for the DeepSeek Harness LLM seam (design-verification twin of dsh-llm-deepseek)

LLM eval & testing toolkit