Compass is Cohere’s retrieval platform for developers building AI applications with their enterprise data. It surfaces the most relevant information from your company’s corpus for use in retrieval-augmented generation (RAG), search, and agentic workflows. Now, Compass is entering private beta as a managed offering - Compass Cloud.
Compass was built for builders. It provides the retrieval foundation that developers can freely configure around their data and workflow requirements. Instead of assembling and operating the stack themselves, teams access Compass through its APIs, MCP server, or Python SDK to shape their desired end user experience.
Until now, Compass has principally powered retrieval for North, Cohere's enterprise agent workspace, including its document libraries and MCP ecosystem. We have also shipped Compass into highly secure, self-hosted environments for partners in regulated industries whose workloads cannot be offloaded to SaaS.
Customer demand for a managed option has been clear and consistent: teams want Compass' best-in-class retrieval capabilities, but many do not want the operational overhead that comes with self-hosting.
We’ve listened to those calls. Compass Cloud expands Compass to a broader market. It lets Cohere manage the full pipeline and model inference, so that our customers can focus even more on building. In parallel, self-hosted deployments remain available for privacy-constrained projects.
We're working with a limited number of enterprise teams as beta partners. Interested? Request access.
Compass is solving open problems in enterprise search
Search and retrieval have improved, but enterprise performance is no longer defined by relevance and latency for a single query. As retrieval becomes key supporting infrastructure for generative AI and agents, three developments are changing the requirements:
- Token economics: Every irrelevant result passed to a model consumes tokens and occupies limited context space. More precise retrieval creates smaller, higher-quality inputs, reducing inference costs and cutting the time needed to complete a task. Retrieval is one of the most effective cost levers available to businesses today.
- Agentic access patterns: Agents may issue dozens of queries while completing one task, reformulating requests and traversing multiple sources. In these multi-hop loops, latency accumulates, relevance can drift, and permissions must be enforced at every step. Retrieval must therefore perform reliably across sequences of machine-generated queries, not only single-shot searches.
- A fragmented retrieval stack: Production pipelines often combine separate systems for ingestion, indexing, reranking, access control, and orchestration. Different middleware and sources of truth leave teams spending substantial effort on integration rather than retrieval quality.
Compass exists to address each of these: 1) by providing relevant, governed context for generative workloads; 2) supporting both agentic and conventional search applications; and 3) consolidating the core retrieval stack into an integrated enterprise platform.
Full-stack search and retrieval
Compass packages document processing and retrieval into one configurable service. Teams can access a single interface instead of integrating and operating separate services.
Connect: Access out-of-the-box connectors for your file sharing and cloud storage workspaces, such as SharePoint, OneDrive, and Google Drive. Quickly access the content you need with Compass’ near-universal data compatibility — multilingual, multimodal, and file format-agnostic.
Parse: Turn complex documents into searchable, structured data. Compass transforms multimodal enterprise files into AI-ready content, applying the right parsing strategy to each document and using vision processing only where it adds value, reducing unnecessary model usage.
Embed: Capture meaning and exact terminology. Compass generates dense and sparse representations together, so search can match both semantic intent and domain-specific language across text and multimodal content.
Index: Keep your source files, parsed content, and embeddings as separate records, so a new embedding model can be adopted without crawling and uploading the same content again. At search time, they sit in one index with their metadata, reducing the need to keep each system in sync.