Solution · Investor Diligence

The deck says one thing. The record shows another.

What a company actually ships, charges, and hires for, dated and sourced from its own public surfaces, against what the deck claims.

Behavioral evidence, not sentiment. Every fact links the page it came from.

The job

Diligence runs on documents the company wrote.

By the time a data room opens, everything in it has been curated. The public record is harder to curate: the pricing page moved, the enterprise tier appeared, the security certifications landed, the job board tilted toward sales. Those are behaviors, and they are dated.

Bixel keeps that record continuously, so the question becomes “what changed, and when,” instead of “who do we believe.”

  • Growth claims with no observable shipping cadence behind them
  • “Enterprise-ready” decks over a site with no trust page
  • Headcount stories the company’s own job board contradicts
  • No dated trail from a claim back to a source

The API

5 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.

GET /v1/companies/{company}
Current structured state across all ten dimensions.
open
GET /v1/companies/{company}/history
Every dated change; the values unlock with a key.
masked · key unmasks
GET /v1/companies/{company}/jobs
Live postings with the comp breakout where stated.
open
GET /v1/companies/{company}/claims
Company-stated claims checked against observation.
free key
GET /v1/companies/{company}/evidence
The capture trail behind any fact, point-in-time.
free key

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

Pull Pinecone’s record: what changed in pricing and hiring over the last year, and which security claims are independently observed?

A real example, live now

Pinecone’s stack page splits observed technology from what job postings state, and marks the overlap verified: see claimed vs observed →