Compass

What it does

Compass takes in enterprise documents and data from connected sources or local uploads, including PDFs, PPTs, DOCX and XLSX files, along with images and spreadsheets. It parses and indexes that content in a managed index, then returns search results and retrieved context for queries from users or applications.

Teams use Compass to give AI agents and retrieval-augmented generation applications access to internal knowledge, and to combine several knowledge stores into one search system. It is built on Cohere's Embed and Rerank retrieval models and includes document-level security and role-based access controls. It can run in a customer's VPC, on-premises, or on Cohere-managed Model Vault, and can connect to an existing stack through Compass MCP.

Compass suits enterprises and developers building search or agent systems over internal data who want a managed index without running their own vector database. Compass Cloud is currently in private beta, so access starts with a request form. Request a quote from Cohere for Compass pricing and deployment terms.

What it can do

• Connects to data sources or accepts uploaded local documents
• Parses PDFs, PPTs, DOCX and XLSX files and processes images and spreadsheets
• Maintains a managed index without requiring a self-hosted vector database
• Uses Embed and Rerank retrieval models for search quality
• Provides role-based access controls and document-level security
• Deployable in a VPC, on-premises environment, or Cohere-managed Model Vault

Works with

• Compass MCP (connect Compass to an existing stack)

Example workflows

Internal knowledge search for an agent

Inputs: company PDFs, slides and spreadsheets connected from a data source or uploaded locally. Steps: Compass parses the files, builds a managed index, and retrieves relevant passages for a user query. Output: ranked search results and retrieved context the agent can use in its answer.

RAG application over consolidated knowledge

Inputs: several internal knowledge stores combined into one Compass search system. Steps: documents are indexed under role-based and document-level access controls, and the RAG application queries Compass for context. Output: retrieved passages that ground the application's generated responses.

Before you choose

• Compass Cloud is in private beta, so access is by request and availability may be limited.
• Request a quote from Cohere for Compass pricing; custom enterprise pricing is offered.
• Deployment is in a VPC, on-premises, or on Cohere-managed Model Vault.

Listed under

• Productivity
• Knowledge and search
• Data connectors
• Developers
• Document parsing
• Document search
• Enterprise search
• Private deployment

Get started

Compass Cloud is in private beta; the page offers a request access action.

1. Request access to the Compass Cloud private beta, or contact Cohere sales for a private deployment.
2. Connect data sources or upload local documents to Compass.
3. Choose a deployment option and connect Compass to your agents or RAG application, optionally through Compass MCP.

Request a Compass quote from Cohere for your deployment. Shared API and Model Vault rate tables do not establish a Compass subscription price.

We couldn’t find published prices for Compass. See their pricing page ↗

Sources for this overview

Product pages, documentation, and pricing checked 28 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Compass: Intelligent Search and Discovery | Cohere ↗cohere.com
• Pricing | Secure and Scalable Enterprise AI | Cohere ↗cohere.com
• Enterprise Open Source AI Models | Cohere ↗cohere.com

Company website

cohere.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Web search

Confirms the official site

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 28 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.