Braintrust:
the business behind the product
Instruments AI traces and evaluation loops with datasets, scores and human review.
Research reviewed 2026-10-22. Financial disclosures below have not been independently audited by ToolJovia. AI's share of the reported result is not independently established.
Put the mechanism to work
Compare two independent reviewer labels under one rubric, list disputes and missing labels, and calculate marginal-adjusted agreement without a quality guarantee.
The task below is an independent ToolJovia adaptation. The workbench does not operate the vendor product or establish customer demand.
Founder or operator
B · founder / official disclosure As of 2026-02-17[b5b-braintrust-series-b]Founder/vendor disclosure only; not independently audited or a typical operator outcome.
What the product does
B · founder / official disclosure[b5b-braintrust-seed] [b5b-braintrust-review]Founder/vendor disclosure only; not independently audited or a typical operator outcome.
Customer and job
B · founder / official disclosure[b5b-braintrust-series-b]Founder/vendor disclosure only; not independently audited or a typical operator outcome.
Why a buyer might pay
B · founder / official disclosure[b5b-braintrust-series-b]Founder/vendor disclosure only; not independently audited or a typical operator outcome.
What is sold
B · founder / official disclosure As of 2026-10-22[b5b-braintrust-pricing]Founder/vendor disclosure only; not independently audited or a typical operator outcome.
Traceable timeline
B · founder / official disclosure[b5b-braintrust-seed] [b5b-braintrust-series-b]- date
- 2023-12-13
- event
- Seed/customer partnership account published
- date
- 2026-02-17
- event
- Series B announcement
Founder/vendor disclosure only; not independently audited or a typical operator outcome.
First customers
B · founder / official disclosure[b5b-braintrust-seed] [b5b-braintrust-series-b]Founder/vendor disclosure only; not independently audited or a typical operator outcome.
Acquisition evidence
B · founder / official disclosure[b5b-braintrust-seed] [b5b-braintrust-series-b]- evidence
- Founder/customer conversations
- Early investor introductions and product partnerships
- current Channel Shares
- Not established
Founder/vendor disclosure only; not independently audited or a typical operator outcome.
AI, API and software stack
B · founder / official disclosure[b5b-braintrust-seed] [b5b-braintrust-review]Founder/vendor disclosure only; not independently audited or a typical operator outcome.
Human work in a proposed comparable pilot
D · unverified / analysis As of 2026-10-22ToolJovia independent analysis or proposed bounded experiment; not founder testimony, verified buyer demand, safety certification or income forecast.
Financial evidence, kept separate
ARR and MRR are recurring run-rate metrics. Revenue is not net profit; financing and valuation are separate from customer earnings. Missing values are unknown, not zero.
| Metric | Reported value / period | Evidence and limitation |
|---|---|---|
| Revenue | Not established | D · unverified / analysis Not established by the primary pages read in this bounded research. |
| MRR | Not established | D · unverified / analysis Not established by the primary pages read in this bounded research. |
| ARR | Not established | D · unverified / analysis Not established by the primary pages read in this bounded research. |
| Gross profit | Not established | D · unverified / analysis Not established by the primary pages read in this bounded research. |
| Net profit | Not established | D · unverified / analysis Not established by the primary pages read in this bounded research. |
| Financing |
| B · founder / official disclosure As of 2026-02-17[b5b-braintrust-series-b] Funding only; not company revenue, ARR, valuation or operator earnings. |
| Valuation | Not established | D · unverified / analysis Not established by the primary pages read in this bounded research. |
What remains unverified
Startup costs; Operating costs. Check the field-level limitations before relying on this case. No missing cost is assumed to be zero.
A task a small operator could test
D · unverified / analysis As of 2026-10-22ToolJovia independent analysis or proposed bounded experiment; not founder testimony, verified buyer demand, safety certification or income forecast.
Working hypothesis about the buyer
An AI quality consultant calibrating two independent reviewers on permitted fixtures
A concrete pilot artifact
AI Review Agreement Calculator report
- Freeze one rubric and obtain independent labels from two distinct reviewer IDs.
- List all expected fixtures, then enter labels without replacing unresolved differences.
- Adjudicate disagreements and missing labels with an accountable expert before using the dataset for acceptance.
What this pilot does not reproduce
The vendor platform, hosted integrations, distribution, real-time execution, rights, customer base and expert review are not reproduced by this browser-local fixture tool.
Still unknown
- Current channel effectiveness
- US buyer willingness to pay
- Full customer acquisition and human delivery cost
- Actual user income/profit
A small, falsifiable pilot
ToolJovia analysis / suggested experiment (D). These steps are not reported founder actions or verified demand.
Mechanism worth testing
D · unverified / analysis As of 2026-10-22ToolJovia independent analysis or proposed bounded experiment; not founder testimony, verified buyer demand, safety certification or income forecast.
Use the fixed synthetic fixture first; then obtain permission for a small client sample, record actual human review minutes and acceptance/rejection, and stop if source evidence or usefulness is inadequate.
Your pilot workflow
D · unverified / analysis As of 2026-10-22- Freeze one rubric and obtain independent labels from two distinct reviewer IDs.
- List all expected fixtures, then enter labels without replacing unresolved differences.
- Adjudicate disagreements and missing labels with an accountable expert before using the dataset for acceptance.
ToolJovia independent analysis or proposed bounded experiment; not founder testimony, verified buyer demand, safety certification or income forecast.
Risks and failure conditions
D · unverified / analysis As of 2026-10-22- Scores are human declarations; the tool does not judge correctness, infer labels or call a model.
- Do not combine different rubric versions or replace missing labels with fail/pass.
- Kappa is undefined when expected agreement is1; no universal good/bad cutoff or hiring/medical decision use.
ToolJovia independent analysis or proposed bounded experiment; not founder testimony, verified buyer demand, safety certification or income forecast.
- Before spending, define an observable buyer action and a deadline in the validation planner.
- Run a small, consented pilot. Record actual review time, rework, replies and payments in your downloaded worksheet.
- Compare observations with your predeclared stop rule. Filling out a plan does not prove demand.
Tools for this mechanism
Original sources and dates
Links support the attributed claims. Sources do not endorse ToolJovia. We do not treat repetition in the press as independent financial verification.
- Braintrust series B
Original founder/vendor/partner publication; B self-reported, not independent financial verification · reviewed 2026-10-22 · published 2026-02-17 · Ankur Goyal original founder account of problem discovery and$80M funding. Funding is not revenue, net profit or a low-cost entry requirement.
- Braintrust seed round
Original founder/vendor/partner publication; B self-reported, not independent financial verification · reviewed 2026-10-22 · published 2023-12-13 · Founder primary disclosure of funding and customer collaborations; customer improvement claims are not independently tested or treated as financial returns.
- How to improve your golden datasets with human review
Original founder/vendor/partner publication; B self-reported, not independent financial verification · reviewed 2026-10-22 · published 2026-05-21 · Official workflow guidance calls for expert expected values, rubrics and calibration. It does not validate a particular human label or universal evaluation rubric.
- Pricing — Braintrust
Original founder/vendor/partner publication; B self-reported, not independent financial verification · reviewed 2026-10-22 · Official free/Pro/custom offers and consumption components; no revenue/MRR/ARR/profit disclosed here.