Resources

Documentation, the roadmap and the words we use.

The documentation carries the how-to guides, reference pages and the decision log, including the pages on the console, the connectors, records and signatures. What changed is generated from the decision log at build time. The claims table below records how we describe an intended platform.

In the documentation

The console, the connectors and the records.

Each page below describes a designed capability and carries its source line. The reference pages are the ones the app and the deck quote.

The console

Connectors

Records and review

Claims discipline

How we describe what is designed, planned and not yet built.

Trovensa would assist scientists, development teams, quality reviewers and manufacturing specialists. It does not claim to independently validate a therapy, authorise its use or guarantee successful development.

Wording we prefer and wording we avoid
PreferAvoid
Surface potential gaps for review.Promising that a platform vouches for a record.
Propose hypotheses with supporting sources.Claiming that agents discover therapies on their own.
Coordinate approved research tools.Suggesting that the research organisation is replaced.
Help teams make better-informed decisions.Suggesting that scientific uncertainty disappears.
Describe the control; name the record it leaves.Claiming compliance with any regulation or standard.

Agent output is agent-generated or proposed. Mock records are labelled illustrative. Every decision stays human and recorded. Timing claims wait for measured data.

What changed

Every decision, dated.

Generated at build time from the decision log. Each entry is an architecture decision record; a superseded decision is never edited, a new one names it.

  1. ADR 0026The public site: preview posture, first-party inquiries and cookieless analytics
  2. ADR 0025Speed is measured, never claimed: event-anchored metrics and baselines
  3. ADR 0024Stage gates: the only writer of the current stage, closed by two signers
  4. ADR 0023Hand-offs as records; event triggers from the audit stream
  5. ADR 0022Decisions waiting for you: inbox items, escalation and delegation limits
  6. ADR 0021Persona twins: bound agent configurations that propose, never decide
  7. ADR 0020Thread visibility, oversight reads, owner-only sharing and conversation retention
  8. ADR 0019Model configuration as data with run-start verification
  9. ADR 0018Program-scoped connections with pinned vendor API versions
  10. ADR 0017Pilot scope, light validation package and status transitions
  11. ADR 0016SOP version binding, QMS authority and impact runs
  12. ADR 0015Read-only connectors, agent service identity and ELN content canonicalisation
  13. ADR 0014ALCOA++ audit envelope, database-computed hashes and an append-only application role
  14. ADR 0013Console protocol: streaming turns, a stateless tool server and per-role run tokens
  15. ADR 0012Public host names: apex for the site, console for the platform
  16. ADR 0011Git hosting on Bitbucket, project TROV, one repository per surface
  17. ADR 0010Repository layout, local repositories without remotes, release deferred
  18. ADR 0009Documentation structure
  19. ADR 0008Research-agent action protocol (no git actions)
  20. ADR 0007Theme tokens from the brand kit, monochrome product accents, 4/8 px geometry
  21. ADR 0006Enumerations as text columns with named CHECK constraints
  22. ADR 0005Identity model with a local development identity mode
  23. ADR 0004Assertion kinds and origin on every finding, hypothesis and proposed step
  24. ADR 0003Program as the tenant scope and the 72-permission catalog
  25. ADR 0002Single brand, product taxonomy, pillars, stages and prefix rule
  26. ADR 0001Adopt the product-forge method and the reference profile

Read on

The documentation carries the reference pages and the decision log.

Open the documentationTalk to us