Skip to content

Guiding principles

TruePPM’s go-to-market is adoption-first: a tool a team must choose voluntarily has to be built for the team before it is built for the org buying the license. Four commitments follow from that, and each one draws a line on purpose. This page is the canonical statement of all four; each links to the page that argues it in full.

Every date and forecast is calculated by a deterministic engine, with the derivation to show for it. A language model may ask your schedule but never invents your dates — when an AI client is connected, the model translates the question and phrases the result, and the engine supplies the number.

The line: an AI-surfaced answer is a computation you can cite, never the model’s opinion. It is one capability with four parts — an agent can compute an answer, cite its derivation, be refused when a change breaks the plan’s rules, and reproduce any answer later — the same contract for a human and an agent.

Computed, not guessed · deep dive: AI-native by design

Logging in through your own identity provider is table stakes for self-hosting, not a paid upsell. Basic OIDC / OAuth2 single sign-on is in the OSS core. The enterprise edition earns its price on identity governance — directory sync, provisioning, enforced org-wide policy — not on the login screen.

The line: log in via your own IdP → OSS; provision, deprovision, and govern accounts from a directory → Enterprise.

SSO is not an enterprise feature

The signals a team generates about its own flow belong to the team. A signal that measures how hard a person — or their agents — is working stays with the team; a signal that answers “will we hit the date” rolls up, as a confidence-weighted forecast, not a scoreboard. Sprint velocity is never auto-exposed to management as a productivity gauge.

The line: team-owned signals roll up only as forecasts, never as per-person scoreboards, and only by the team’s explicit, audited opt-in.

Team ownership is not surveillance

TruePPM is self-hosted and Apache 2.0. It runs on your infrastructure — including air-gapped and regulated environments — and the scheduling engine even ships standalone on PyPI, usable without the API if you only need the math. You are never on someone else’s cloud by default.

The line: the community edition is fully functional on its own and never depends on the enterprise repo; the dependency is strictly one-way.

→ Deep dives: Deployment · AI-native by design §5 · License

A tool a team must adopt voluntarily has to be built for the team first. A signal that becomes a management weapon is a signal the team stops trusting — and an untrusted signal is worthless to management too. A number the model made up is worthless to everyone. The four commitments are one stance seen from four sides: earn adoption by being correct, controllable, and on the team’s side — then let governance be what the organization adds on top.

CommitmentThe line it drawsWhere it’s argued
Computed, not guessedAn AI answer is a citable computation, not the model’s opinionComputed, not guessed
Adoption over gatekeepingLog in via your own IdP → OSS; govern accounts from a directory → EnterpriseSSO is not an enterprise feature
Built for the teamTeam signals roll up as forecasts, never as scoreboardsTeam ownership is not surveillance
Your data, your infrastructureSelf-hosted, Apache 2.0, one-way dependency on enterpriseDeployment

New here? Why now puts these commitments against the current market shift; The Story walks a hybrid program end to end.