3 projects
aither-adk
Build AI agent fleets. Any backend, local or cloud. Effort-based routing, 48 identities, knowledge graph memory.
aither-kvcache
Near-optimal KV cache compression for LLM inference — TurboQuant (vector quantization) + TriAttention (spectral compression)
aithersdk
Python SDK for AitherOS — the canonical client library
