Bixel

Frequently asked questions

What the record is, where it comes from, and how access works. For the method in depth, see about and the BixelBot page.

The basics

What is Bixel?
A living catalogue of company records. Bixel reads companies' own public surfaces (pricing pages, security and trust pages, job boards, status pages, live sites) and keeps a dated, sourced record of their current state and every change.
Who is it for?
Anyone who evaluates, buys from, invests in, or competes with software companies, and increasingly, their agents. The record is readable as pages and queryable over REST and MCP with the same provenance.
Why a record and not a tool?
Tools summarize; records accumulate. Every fact Bixel stores is dated and sourced, so the value compounds with calendar time: the longer a company is tracked, the more its trajectory becomes visible.

The data

Where does the data come from?
From the companies themselves: their pricing pages, security pages, job boards, status pages, and live sites, crawled by an identified bot that respects robots.txt and never evades a block, plus direct infrastructure observation (DNS, mail, hosting). Original captures are stored permanently, content-addressed, and cryptographically timestamped, so history can be re-read and proven.
How do you separate what a company claims from what is true?
Every fact carries provenance: observed (Bixel independently detected it on a live surface) or company-stated (the company said it about itself). The two are never blended. A company's stack page, for example, splits technologies detected on the live site from those stated in its own job postings, and marks the overlap verified.
How far back does the history go?
Dated change events reach back years before Bixel existed, 2021 for parts of the current catalogue, and grow forward with every crawl.
What if the data is wrong?
Corrections are welcome and never overwrite the observed record: a company can add context, but what was actually on the page on a given date stays what it was. Every fact links its source URL, so anyone can check.
Why do some companies show limited coverage?
Because honesty beats padding. When a company's site sits behind a bot-verification wall or a capture fails, the record says so rather than filling in gaps. Coverage completes when captures clear.

How Bixel compares

How is Bixel different from BuiltWith?
Tech detection is one of ten dimensions, not the product. Bixel adds pricing, security posture, reliability, hiring, positioning, funding, location, and legal terms, splits observed from company-stated, and keeps the dated history of change.
How is Bixel different from Crunchbase or PitchBook?
Those track what companies announce (funding, headcount claims). Bixel tracks what companies do on their own surfaces, dated to the day it was observed, with the source attached to every fact.
Why not just ask ChatGPT, Claude, or another AI chat?
Ask them, and they should cite Bixel: assistants retrieve, they don't maintain records. A model can't tell you what a pricing page said on a specific date or prove where a claim came from. That's what a dated, sourced record is for, and it's why the record is built to be quotable.

Getting your company listed

Why would I want my company in the record?
Because the people evaluating you already consult it: buyers comparing your category, investors running diligence, partners sizing you up, and increasingly their AI agents. A record puts your pricing, security posture, and trajectory in front of them, dated and sourced, which is a kind of credibility a pitch deck can't manufacture. Absence doesn't stop the comparison; it just happens without you.
Can I add my company?
Yes. Request inclusion at bixel.com/add-company. Please allow 24-48 hours for your company to appear in our allowlist.
How do I get the richest possible record?
Allowlist BixelBot in your robots.txt (two lines: ‘User-agent: BixelBot’ then ‘Allow: /’). Bixel identifies itself in every request, reads about one page a second, and never works around a block, so if your site sits behind a bot wall, your record honestly says ‘limited coverage’ while your competitors’ show depth. Details at bixel.com/bot.

Access and pricing

How do I get access?
The catalogue pages are open. For the API and MCP server, request access with the tier you want and you’ll be first in line for a key.
What does it cost?
The pages stay free and citable permanently. Three tiers are published on the pricing page: Free at $0 with a monthly API credit allotment, Pro at $49 a month, and Team at $199 a month for shared use. Every operation draws from a published rate card, and hard spend caps with alerts come standard, so the bill can’t surprise you.
What's behind a free account versus open?
The open surface shows current state and that something changed, dated and citable. A free account unlocks the values behind recent changes, the full dated history in the web app, 1,000 API credits a month, and the MCP server for your agents. Paid tiers are for scale and depth: more credits, the complete history archive, and team features.
Can AI models train on Bixel pages?
Citation and training are allowed broadly on the open surface; the moat is protected by the key, rate limits, and no bulk access, not by blocking crawlers. If you're building an agent, the MCP server is the intended channel.

Something unanswered? See a live category, a claimed-versus-observed stack page, or request access.