AI Qadam Build is the foundation under our four streams (Events, Education, People, Accelerator). Open-source software the community owns and runs itself — so the ecosystem stays localised, independent, and free of vendor lock-in.
Project stages are defined in GOVERNANCE §8 — Incoming (proposed or adopted, has an owner and a permissive license), Active (has a maintainer, contributors beyond the founder, docs, writes to the graph), Graduated (sustained activity, multiple maintainers, real deployments).
Design and brand infrastructure for the AI Qadam umbrella — tokens, components, identity, voice, merch templates. Every Build project (and chapter site) loads its CSS and SVGs from here live, so the umbrella stays drift-free without a mirror-and-commit step.
Open-source workflow automation. Forked from the MIT core of Activepieces with enterprise code removed. Localisation and regional integrations are the current focus. The first Build project.
Got something the region needs? Build projects come two ways (§3): adopt and localise an existing open-source base, or build new from zero. Open an issue with the problem it solves, the stream(s) that need it, and who would maintain it — AI Qadam Global decides what enters Incoming.
Build projects come two ways, both first-class. The test is simple: if a good open-source base exists that we can localise, prefer it. Build from zero only when nothing fits, or when adopting would compromise the mission (license, data residency, language).
Fork an MIT/Apache-2.0 codebase, strip what we don't need, add what the region does need. Cheaper, faster, lower risk. Honour upstream attribution. The Qadam Flow path — forked from Activepieces.
When there is no good base to fork, or when adopting would compromise the mission, build from zero — but only with maintainers ready to carry it. Avoid reinventing wheels for the sake of it, and avoid over-scoping. The reputation graph path.
Build projects are a living testbed for agentic development. Routine maintenance — piece scaffolding, localisation, generated tests — runs through AI agents. Maintainers hold the standard; no agent ships to main alone.
The canonical, timeless rules live in GOVERNANCE.md. The headline rules:
Repositories, trademarks, and domains belong to AI Qadam (Global). Using a project does not confer ownership of it.
Permissive license (MIT / Apache-2.0). No open-core paywalls. SSO, RBAC, audit logs are table stakes, not premium tiers. Contributions use the DCO — contributors retain copyright; no CLA, no copyright assignment.
Partner funding pays for shared infrastructure (hosting, CI, security audits, events, maintainer support). Technical direction is decided by maintainers on merit — never by who paid.
Standing is earned by contribution, not by title: contributor → committer → maintainer. Roles and contributions are recorded in the AI Qadam reputation graph as the merit ledger.
Decisions today sit with the AI Qadam core while the first projects find their feet. A neutral, multi-maintainer foundation — in the spirit of ASF / CNCF, with neutral ownership, money separated from control, and a project-maturity pipeline — is the destination, not the current state.
This document and the structure around it will be distributed and formalised as projects and maintainers from multiple organisations accumulate. We say that plainly rather than claim a structure we have not yet earned.
See GOVERNANCE §10 for the canonical statement.
