Why this matters

The region runs its automation on someone else's infrastructure.

Billing in dollars, servers abroad, interfaces in foreign languages, and the features that actually matter — SSO, RBAC, audit logs — locked behind an enterprise paywall.

Qadam Flow is the region's own platform: self-hosted, data stays home, the locks are off.

This isn't about "free Zapier" — it's about independence.

01

Your data stays with you. Self-hosted deployment — on your infrastructure, in your jurisdiction.

02

No enterprise paywall. The MIT core only — no edition gating, no paid tiers, no enterprise feature flags.

03

Multilingual by default. 11 UI locales today — English, Russian, and 9 others. Uzbek and Kazakh are next, contributors welcome.

04

Regional integrations next. Payme, Click, Uzum, Kaspi — planned, contributors welcome. The framework is modular.

Templates

252 ready-made templates, with the math.

Pick from the library — every template ships with the full step graph, a setup guide, and measured time and cost savings. Or start from a blank canvas.

Browse the catalogue

Daily Schedule & Task Briefing template — saves $1,260/year and 46 hours, with a six-step flow preview on the right (Every Day → weather → calendar → tasks → AI summary). — Daily Schedule & Task Briefing — six steps, end to end. Weather, calendar, tasks, AI summary.

The builder

Drag, drop, run. Webhook in, local LLM out.

A real flow: catch a webhook, send the prompt to a local Ollama (qwen2.5:7b), return structured JSON. No OAuth, no API key, no cloud LLM — the AI step doesn't have to leave your machine.

Flow builder showing 'Ollama AI Webhook' — Catch Webhook trigger → HTTP step 'Ask Ollama (qwen2.5)' → Return Response, with the trigger's Live URL panel open on the right. — Ollama AI Webhook — Catch Webhook → Ask Ollama (qwen2.5) → Return Response. Live URL panel open.

Steps picker searching for 'stripe' — returns the Stripe qadam with triggers Checkout Session Completed, New Payment, New Customer. — Stripe qadam — Checkout Session Completed, New Payment, New Customer. Concrete event semantics.

Integrations

238 qadams. Real triggers, real events.

27 core + 211 community qadams inherited from the Activepieces ecosystem and free of enterprise gating. Not generic HTTP placeholders — concrete triggers like checkout.session.completed and typed actions for Slack, Notion, Postgres, S3, Stripe, and 200+ others.

238

qadams (27 + 211)

252

templates

11

UI locales

MIT

licence

AI

AI is a step, not a feature.

Bring your own keys — Anthropic, AWS Bedrock, Azure, Cloudflare AI Gateway, Google Gemini, Mistral, or anything OpenAI-compatible (Ollama, vLLM, LM Studio). And every flow becomes an MCP tool — plug Qadam Flow into Cursor, Windsurf, or Claude Desktop and your automations show up natively.

AI Providers settings page — Anthropic, AWS Bedrock, Azure, Cloudflare AI Gateway, Google Gemini, Mistral AI — each with an Enable button. — AI Providers — pick your engine per flow, or per step.

Platform MCP Server settings — Server URL 'http://localhost:8080/mcp/platform', Connection and Tools tab switcher. — Platform MCP server — every flow exposed as an MCP tool.

Execution

Every run, every step — inspectable.

Each run lands in a queryable history. Drill into any step to see what it received, what it returned, and how long it took. Retry, replay, debug — the run log is the source of truth, not your guesswork.

Run detail for 'Ollama AI Webhook' — three steps all Succeeded (Catch Webhook 0 ms → Ask Ollama qwen2.5 13 s → Return Response), with the trigger output panel open showing POST method, headers, body, and query params. — Run detail — three steps Succeeded, trigger output expanded inline.

Platform Admin Users page listing demo@aiqadam.org as an Admin, with columns Identity, Name, Role, Created, Last Active, Status. — Platform admin — users, roles, projects. Multi-tenant from day one.

Self-host

One command. Your infrastructure.

Platform → projects → users. Strong tenant isolation enforced at the query level. Runs on your servers, in your jurisdiction.

$ curl -fsSL https://flow.aiqadam.org/run.sh | sh
# brings up postgres + redis + app + workers on port 8080

Full setup guide

Status

Early stage — and honest about it.

What's done, what's in progress

The MIT core of Activepieces has been forked and stripped of all enterprise code. 238 qadams and 252 templates ship today.

SSO, RBAC, and audit logs are scaffolded but not production-ready — they get built as real deployments need them, not on a fixed schedule. Don't run security-critical workloads on Qadam Flow yet. Details in the Roadmap.

Regional integrations (Payme, Click, Uzum, Kaspi) and Uzbek / Kazakh UI translations are planned but not yet built. Contributors are exactly where the project is leaning.

Maintained by AI Qadam — under the founder's stewardship for now, but structured to belong to the community (DCO, no copyright assignment).

How to join

You don't have to be an expert — caring is enough.

A simple, meaningful start:

Build a qadam

Write the integration the region needs — Payme, Click, Uzum, Kaspi, or any local API. About a day of work with the qadam-builder skill.

Translate the UI

Uzbek and Kazakh are missing. No code required — translation strings live in JSON. Russian and 9 other locales are already shipped and can be polished.

Test and report

Run it, break it, file an issue. A reproducible bug is half the fix. Self-hosters who report what doesn't work are gold.

How to start AI Qadam

Workflow automation — built by the region, for the region.