Candour
What is not built yet
Draft. States as recorded in the repository on 2026-10-13. Nobody re-ran the checks for this draft.
Cautious organisations distrust sales language, so this list has a page of its own, and every page links to it. Each row names its issue, because a gap is not recorded until it is an issue.
The first things to know
- The planner, the five standard reviewers, feedback from test users and the five measures are planned and not built.
- The board and funder view is planned and not built.
- The compliance records are mostly planned, and B Team holds no certificate.
- The test-set gate is a command a person runs, not yet enforced by the system.
- The European route and the hosted service are intended and not connected.
- Security is not complete today: use AI and stay compliant lists what is not met.
- It is an early build, and it is not offered to anyone yet.
Where other tools are ahead today
A cautious buyer finds these in a minute, so we say them first. The first is the one an organisation's product owner is most likely to try instead: Replit. Each is the vendor's own page, as a tool returned it (tool summary), read on 2026-10-13. A vendor's page describes its product. It does not prove it works.
| Where | What the pages say |
|---|---|
| Idea to a hosted app, for people who do not code (Replit) | Replit is what an organisation's product owner is most likely to try instead: "Tell Replit Agent your app or website idea, and it builds it through chat". Its Agent builds apps from chat and tests them in a browser, and Replit hosts and deploys the result. Agent 4 is said to split a request into tasks that run in parallel and merge after approval (search summary, read 2026-10-12). For someone who does not code, Replit is faster from an idea to a running app than B Team, which has no hosting yet. Replit runs on Google Cloud, region not named. The pages read name no models, no non-profit offer, no backlog it manages and no human-approval step before deploy. |
| Discovery and feedback | Productboard has more than 25 integrations, importance scoring, portals and citations on its findings. B Team's feedback is new and not built. |
| The agent engine and its ecosystem | Developers compare Claude Code, Cursor and Codex. They ship years of work on agents, sandboxes and approvals. |
| Certificates | Replit names SOC 2 Type II and GDPR, and is working toward ISO 27001. Cursor and Productboard name SOC 2 Type II, ISO 27001 and ISO 42001. Anthropic publishes its report through a trust centre. B Team holds none. |
| Hosting in Europe | Lovable, Linear, Copilot (GHE.com) and Productboard (in beta) offer EU regions. B Team's is later. |
| GitHub | Replit names GitHub import (GitLab not found in the docs read). Productboard, Copilot and Claude Code cloud sessions work with GitHub. B Team's GitHub support is later. |
| Non-profit prices on tools | A Claude Team seat at $8 for non-profits, Jira and Confluence free up to 25 seats, GitHub Team free, GitLab Ultimate free up to 20 seats. These cover tools, not the whole service. |
The limits, each with its issue
| Limit | Tracked as |
|---|---|
| The reviewer is the same model family as the worker. | #82 |
| Temperature cannot be set. The test sets test the duty not to invent. | #25 |
| There is no fallback model. Failover did not fire on a provider 404. | No issue found yet |
| The test-set gate is a person's command. The continuous-integration system holds no provider key. | #115 |
| The prompt-injection suite measures. It does not certify. | No issue found yet |
| Nobody has shown that a reviewer catches a real fault. | #28 |
| Early build: only two projects are registered, both on this repository. | No issue |
| Keys sit in plain files. Traces hold full words. | #51, #84, #57 |
| One provider key and one bill across projects. | #170, #171 |
| No data policy is checked against any provider's terms. | #57 |
| No audit log. No secret scanning in the build pipeline. | #56, #116 |
| No sandbox for code that agents run. | #176 |
| The stop switch halts spending, not a job that is running. | No issue found yet |
| Approval is as strong as the reviewer's GitLab login. | No issue found yet |
| There is no single queue across projects. | #111, #31 |
| The web page cannot start or stop anything, and has no sign-in. | #113 |
| It runs with Docker Compose on a laptop. The server deployment is not built. | #44 |
| The planner, the register, standard reviewers, feedback and measures are proposals. | #286, #287, #288, #290, #299 |
| No compliance reviewer yet. Its first rules would be the GDPR and the AI Act. | #107 |
| No record of each change, no model link on it, no evidence pack (change record, model provenance, evidence pack). | #302, #307 (claims 62 to 64) |
| No supplier package for your buyer, and no incident notice (supplier package). | #309 (claim 65) |
| No processing agreement template. B Team's access to client data is not yet off by default (processing agreement). | #297, #300 (claim 66) |
| B Team holds no certificate. Independent assurance has no plan with dates (independent assurance in stages). | #306 (claim 67) |
| B Team's own screens are not tested for accessibility (accessibility of our screens). | #311 (claim 72) |
| No standard agent ships its own test set, and the gate is not enforced by the system. | #115; #316 (claim 69) |
| No published contact for when something breaks, no vulnerability policy, no reporting desk. | #303 |
| A preview of a change before you accept it is planned. It needs the sandbox first. | #176; #292 |
| The hosted service in Europe is intended. Its preconditions are not met. | #295 |
| The Scaleway key answers but is not connected to the engine, and it is replaced first. | #81; #294 |
| Open models are not yet measured against a top commercial model. | #314 |
| The board or funder view is planned. | #293 |
| The licence is not checked by a lawyer. Who owns what the agents write is unsettled. | #22 |
| The name B Team is not checked as free to use. | #317 |
Where a row says "No issue found yet", an issue is needed before this site can publish the row. The model versions in use are pinned, and some retire within twelve months of 2026-09-26 (watched in issue #117).
Everything in the ledger that is not built
42 of the 73 claims in the ledger are not built. This site calls them "planned", and never says the product does them. Each number links to its row on the evidence page.
| No. | Planned, not built | Tracked as |
|---|---|---|
| 7 | Limits on steps, tokens and wall time per job. | #58 |
| 12 | The reviewer is a different model family from the worker. | #82 |
| 15 | Named reviewers per project and kind of work. | #111 |
| 16 | Review items show figures, quotations and preserved sources. | #40 |
| 25 | Secrets stay out of prompts, traces and logs. | #51, #84, #57 |
| 26 | Code agents run in sandboxes apart from the host. | #176 |
| 27 | A full, tamper-evident audit log. | #56 |
| 28 | A data policy says which provider may see a project's data. | #57 |
| 30 | Start and stop work from the page. | #113 |
| 38 | The engine improves itself: evaluation, and model swaps proposed. | #35 to #39 |
| 39 | A project adds checks and approvers that only raise the bar. | #198, #199 |
| 40 | Research-to-publish with sources and a publication gate. | #40 |
| 42 | Software, infrastructure and dependency agents. | #10, #44, #46 |
| 45 | A planner proposes safe parallel work. You can change it. A register stops clashes. | #286 |
| 46 | Five standard reviewers, one per kind of review: security, documentation, compliance, Done when, architecture. | #107; #299 |
| 47 | Your team chooses a human review level, shown on every change. Level A, with the line "no human engineer has read this code", is decided. | Level A decided 2026-10-13. #299 |
| 48 | A goal becomes an issue with a Done when, drafted by the engine and approved by you. | #287 |
| 49 | Test users answer a link and a few questions without any account. | #288 |
| 50 | The code host is the single record, behind one layer. Self-managed GitLab first. | #279; #289 |
| 51 | Each task runs in real time or in a cheaper batch. | #196, #65 |
| 52 | Five measures are recorded by the machinery and shown together. | #290 |
| 53 | A guided setup into your own account or server. | #291 |
| 54 | You may pick only models your data policy allows that passed the test set. | #294 |
| 55 | A hosted service in Europe separates each organisation. Intended only. | #295 |
| 56 | Nothing requires a person to answer in real time. Waits have a stated turn. | #65, #196 |
| 57 | Each security requirement's state is shown in your workspace. | #113, #154 |
| 58 | Standard agents, instructions and checks come ready-made. A team adjusts them. | #318 |
| 59 | Models run in Europe in steps: reviewers on Scaleway, then all models, then hosting. The Scaleway key answered 200 and listed 16 open models on 2026-10-04. It is not connected to the engine, and it is replaced (#81) before it is. | Decided 2026-10-13. Not built. #294 |
| 60 | A role moves to an open model only after a test-set run against a top commercial model. The nearest built part is the test-set gate (claim 20). | Decided 2026-10-13. Not built. #314 |
| 61 | A shareable view for a board or funder: what was asked, built, accepted and what it cost. | New requirement agreed 2026-10-13. #293 |
| 62 | Each change has a record: diff, test result, findings, risk note with rollback, approver, time. | #307 |
| 63 | Each change records which model, provider and version wrote or reviewed it, and the agent role. | #302 |
| 64 | You can export an evidence pack for a period, in open formats. | #307 |
| 65 | A package for your buyer or auditor, from live data: suppliers, data flows, exit plan. | #309 |
| 66 | A processing agreement before B Team sees personal data. No own use. Its access off by default. | #297, #300 |
| 67 | Independent assurance for B Team in stages, with dates. B Team holds no certificate today. | Waits for a decision by the founder (Codrin). #306 |
| 68 | Standard agents build no donor-profiling features without a named person and an impact assessment. | #298 |
| 69 | Each standard agent ships a test set. The system, not a person, enforces the gate. | #115; #316 |
| 70 | Code checks what code can. AI-written checks run without AI. Each fault becomes a test. | #296 |
| 71 | You can read each agent's test results and traces in plain words. | #113, #154 |
| 72 | B Team's own screens meet WCAG 2.1 level AA, with a statement. Pages the agents build get a check. | #311 |
| 73 | Short training material on what the agents can and cannot do, for your staff. | #302 |
Claims 59 to 73 come from decisions and research of 2026-10-13 and need an issue each before the site can publish them. Planned, not built