Agent grounding over MCP
Connect the Bixel MCP server and company questions get answered from a dated record with the capture behind every value, instead of from whatever the model absorbed in training.
Fourteen tools over one connection, available on the free tier.
The job
A confident answer and a correct answer look identical.
Ask any assistant what a private company charges, who it integrates with, or whether it holds a security certification, and you get fluent prose. Some of it is right. None of it carries a date or a source, so the reader cannot tell which parts to trust, and neither can the agent.
Over MCP the same question resolves against structured facts that each name their as_of date and the page they were read from. The agent quotes the record, cites it, and can hand back the capture when challenged.
What that costs today
- Answers about private companies with no date attached
- Training data that froze before the pricing page changed
- Citations to a homepage that never stated the fact
- No way for the reader to check the agent's work
The API
4 calls cover the job.
Current state is open. History values, claim checks, and evidence trails sit behind a free key. These are the real routes.
/v1 /companies /{company}The whole current record in one call, behind get_company.
/v1 /companies /{company} /factsOne dimension at a time, each fact carrying provenance and date.
/v1 /searchResolve a company name to the domain the record keys on.
/v1 /companies /{company} /facts /{key} /proofThe portable proof bundle for a single fact.
For agents · MCP
Or let your agent run it.
Connect the Bixel MCP server and any agent, Claude, an internal copilot, a procurement bot, works this job against live structured data, with every value citing the dated capture it came from. The recommendation becomes auditable, not a guess.
You ask your agent
“What does Pinecone run in its stack, and when was each of those observed?”
Your agent stops guessing about companies.
The MCP page lists every tool, its credit cost, and whether it needs a key: see the tool list →
More in AI & Agents
The answer layer for agents that need company facts with receipts.
Claim verification
Given a generated sentence about a company, return supporting or contradicting facts with provenance.
Explore
RAG corpora
Licensed slices of the record, by category or dimension, as retrieval sets for vertical AI.
Explore
Factuality benchmarks
Dated ground truth for evaluating models on what was true as of a given date.
Explore
Entity resolution
Domain to company to parent, with subsidiaries and renames handled.
Explore
Procurement agents
Agents that shortlist, compare, and monitor vendors use the record as their memory.
Explore