Connect
Write documents, upload files or connect an app. Each lands in a namespace.
Search across 100M+ documents
Conare finds the right answer in any data you bring, for your agents and the products you build on them.
Preparing the Northwind close plan
Searching 6,240,000 documents3 of 6,240,000 documents
call recordings · emails · CRM records · proposals · signed contracts
Before they sign
Finding investors for your seed round
Searching 112,480,000 documents3 of 112,480,000 documents
SEC filings · scanned pitch decks · cap tables · fund letters · news
Shortlist
When does the new checkout go live?
Searching 9,640,000 documents2 of 9,640,000 documents
issues · pull requests · code · Slack threads · design docs
Claude
Thursday 12 November,Linear as long as the payment tests pass the day before.Slack
Can I switch to yearly billing halfway through a month?
Searching 2,418,000 documents2 of 2,418,000 documents
tickets · help articles · chat transcripts · screenshots · release notes
Help agent
Yes. The switch happens straight away,Notion and the unused days come off your first yearly invoice.HubSpot
Reviewing the Stripe contract before renewal
Searching 1,260,000 documents3 of 1,260,000 documents
scanned contracts · PDFs · redlines · emails · CRM records
Terms to know
Integrations
How it works
Conare chunks, embeds and indexes on write, keeps your data updated in real time, and reranks every query.
See it on your own data.
Read docsWrite documents, upload files or connect an app. Each lands in a namespace.
A changed record replaces its old version. A deleted one is deleted.
Keyword and vector search find candidates. The reranker puts the best first.
Your model gets the passages that answer the question, with their source.
Products
Store documents and query them.
Keep your apps in a namespace.
Read any file in full, every row and cell.
Embeddings for any vector database.
Put the best passages first.
GDPRCertified
Personal data handled to the EU's data protection rules, with data-subject rights honored.

SOC 2 Type IICompliant
An independent audit of how our security controls operated over three months.
Private deploymentAvailable
Search and the models behind it, running inside your own cloud account or data center.
Parse
Spreadsheets, workbooks and PDFs, every row and cell, ready for your agents to search.
Read docsBenchmarks
Conare is 1st on all 8 benchmarks.
See the full benchmarks89.4%
LegalBench-RAG, answered right. 1st of 4.
85.4%
LegalBench-RAG, Recall@5. 1st of 9.
96.2%
NEEDLE Scholar (arXiv), Recall@10. 1st of 20.
93.9%
NEEDLE Legal, Recall@5. 1st of 21.
95.0
ViDoRe v3, RCP-nDCG. 1st of 17.
89.2
BRIGHT, RCP-nDCG. 1st of 17.
90.3
NanoBEIR, RCP-nDCG. 1st of 17.
75.0
LMEB, nDCG@10. 1st of 108.
Data from 13 October to 10 November 2026. Conare through its live API. Legal rows are our runs; the NEEDLE rows are what NEEDLE published; on the RCP-nDCG rows the AI-judged labels and the other scores are Cohere's open release, and OpenAI's and NVIDIA's are our runs. Every benchmark in full →
Price vs accuracy
On ViDoRe v3, conare-embed-1 scores higher than every embedding API on this chart and costs less per token than each of them.
Price per 1M tokens, log scale
List prices per 1M tokens. Mixedbread publishes no embedding price; its $1.50 is indexing per 1M content tokens. Perplexity's open 9B has no API price, so it isn't plotted.
FAQ
Conare is a retrieval API for AI agents and RAG apps. It stores your data, keeps it in sync and returns the right context to your AI, through the API, the SDK or MCP.
Conare is invite-only for now. Book a demo: we set up your organization with you, and your team then creates API keys in the console.
The platform when you want Conare to hold your data, keep it in sync and return the right context. The models or the vector database alone when you already run the rest of the stack.
No. We do not train models on your content, unless your organization asks us in writing for a custom model built from it, which then serves your organization alone.
On Microsoft Azure and Google Cloud in the United States. Every vendor that processes customer data is on our subprocessors page.
A changed record replaces its old version. A deleted one is deleted.
One call removes everything Conare holds of them.
Nothing. Conare runs the servers, shards and index tuning, the embedding and reranking models, and the sync jobs for every app, and keeps each organization's data separate.