Search

Adapter LangChain.js do Riv — tools prontas (riv_authorize, riv_get_activity) e guard determinístico para tools de pagamento.

Type-only package that declares the OpenAI Codex provider identity for the Makaio framework. It exports a single `providerDefinition: ProviderDefinitionInput` for the Codex App-Server. Codex is a subprocess-based provider — it communicates through the Cod

Talk to your files. Local RAG that actually works — llama.cpp + turbovec.

Pack a codebase into one LLM-ready Markdown file with a file tree and token estimate. Zero dependencies, no server.

Python SDK for CollieAi — safe customer-owned LLM streaming and input moderation.

Expo native module for TidGi-Mobile: filesystem I/O + TiddlyWiki .tid/.meta/.json batch parsing + git status in Kotlin (Android) and Swift (iOS)

Unified multi-provider AI image generation library (Gemini, OpenAI, xAI)

Python SDK for Loupe — observability for LLM apps (latency, tokens, cost, errors).

Robust extraction of structured payloads from LLM output via sentinel blocks (<<<TAG>>> … <<<END>>>).

FastAPI backend scaffold — user auth, social, subscriptions, jobs & LLM

BullMQ (Redis) Transport for nestjs-durable — cross-process and cross-language steps

AI observability & debugging — diagnose why LLM outputs fail and get specific fixes.

Boruna: a deterministic, capability-safe programming language and framework for LLM-native applications