Search

Code avec les modeles GPT, Claude, Gemini, DeepSeek et Grok, avec des reglages, dans un panneau a cote de ton code.

Temporal weighted average for VapourSynth with no cap on the number of weights

HuRI module: Retrieval-augmented answers with long-term memory (Qdrant + an OpenAI-compatible LLM), and the question aggregator

Placeholder that sets up trusted publishing. The first release of @miragon/pulse-camunda7-client is 0.21.0.

Temporary package placeholder for staged publishing

Small-model classifier for dsh-tool-manager: places tool groups into categories once, caches the answer, and never classifies on the hot path.

LangChain.js tools for AgentSearch: free live web search, plus pay-per-call extract and JS render ($0.005 USDC on Base via x402)

OpenAI Realtime adapter for vRealTime, plus HTTP speech-to-text and text-to-speech for cascade

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