Morph
Morph builds fast models that improve AI coding agents. Its OpenAI-compatible API serves the Apply (Fast Apply) model that deterministically merges LLM-generated code edits into source files at 10,500+ tokens/second, plus code embeddings and reranking. Morph Cloud adds Infinibranch microVM sandboxes that snapshot, branch, and restore entire VM states in under 250ms for agent workloads.
APIs
Morph Apply API
OpenAI-compatible chat-completions endpoint that merges an LLM's update snippet into the original source with structure awareness, using the morph-v3-fast, morph-v3-large, and a...
Morph Embeddings API
OpenAI-compatible embeddings endpoint producing vectors tuned for code and retrieval, using the morph-embedding-v4 model with float or base64 encoding.
Morph Rerank API
Cohere-client-compatible reranking endpoint that scores documents (or embedding IDs) against a query for code-aware retrieval, using the morph-rerank-v3 model with optional top_...
Morph Cloud Sandboxes API
Infinibranch microVM (Firecracker) sandbox infrastructure for agents - a user-scoped API for managing instances and snapshots with sub-250ms branch, snapshot, and restore, comma...