Compact
Context-compression API that shrinks a coding agent's chat history or code context by 50-70% at 33,000 tokens per second while keeping every surviving sentence verbatim, at $0.20 per million input tokens and $0.50 per million output tokens.
Sources
- Pricing
-
Page text fetched 2026-11-16 (curl, full text read); page quotes: "for $0.20/M input and $0.50/M output"; "Free tier available. Pay only for what you use.".
- Description
-
Official documentation · 16 Nov 2026
Page text fetched 2026-11-16 (curl, full text read); page quote: "An LLM context compression API: shrink context 50-70% at 33,000 tok/s while keeping every surviving sentence verbatim, for $0.20/M input and $0.50/M output.".
- What it does
-
Official documentation · 16 Nov 2026
Page text fetched 2026-11-16 (curl, full text read); page quote: "An LLM context compression API: shrink context 50-70% at 33,000 tok/s while keeping every surviving sentence verbatim, for $0.20/M input and $0.50/M output.". No capabilities term describes verbatim context compression (summarization is a faithful reduction, not a line-for-line cut), so capabilities stays empty rather than take the nearest wrong term.
- Sold within
-
Page text fetched 2026-11-16 (curl, full text read); page quotes: "Get Started"; "for $0.20/M input and $0.50/M output". Self-serve sign-up and a published per-token rate on the product page; no parent offering is named.
- Name
-
Official documentation · 16 Nov 2026
Page text fetched 2026-11-16 (curl, full text read); page quote: "An LLM context compression API: shrink context 50-70% at 33,000 tok/s while keeping every surviving sentence verbatim, for $0.20/M input and $0.50/M output.".
- URL
-
Official documentation · 16 Nov 2026
Page text fetched 2026-11-16 (curl, full text read); page quote: "An LLM context compression API: shrink context 50-70% at 33,000 tok/s while keeping every surviving sentence verbatim, for $0.20/M input and $0.50/M output.".
- Interfaces
-
Official documentation · 16 Nov 2026
Page text fetched 2026-11-16 (curl, full text read); page quote: "Read the docs". The docs index lists POST /v1/compact and a TypeScript SDK call for Compact.
Also from Morph
-
Morph Fast Apply Model API
Specialised model API that merges AI-generated code edits into source files at 10,500+ tokens per second with 98% merge accuracy, in real-time and batched modes, with a self-hosted option.
-
WarpGrep API service
Code-search agent that searches a codebase in a separate context window, running grep, read and list-directory calls in parallel without embeddings, and returns only the relevant code to the coding agent that called it.
-
Reflex API service
Decision API that takes a state (text, JSON or up to 16 images) and typed choice, score and yes-or-no questions and returns a probability for every answer with zero output tokens, at $0.042 per million input tokens.
-
Model Router API service
Prompt classifier that scores a prompt's difficulty, ambiguity and domain in about 180 ms and tells the caller which model to send it to, at $0.005 per request.
-
Dedicated Inference Infrastructure service
Reserved B200 and B300 GPU capacity that serves one open-weight model on an isolated, OpenAI-compatible endpoint with zero data retention and a 99.9% monthly availability SLA, billed by the GPU-hour and invoiced monthly upfront.
Something wrong here? Send a correction — quote this product id: morph-compact.