Autonomous SDLC from Oxytal AI Lab
An AI-powered pipeline platform that takes requirements from Confluence all the way to a reviewed GitHub PR — automatically, auditably, with 8 specialised agents.
The coordination problem, solved.
Every software team runs the same cycle — requirements, architecture, development, review, testing, deployment, monitoring. The steps are well understood. The delays are not caused by the engineering. They're caused by the coordination around it.
ForgePipeline is an AI platform built by Oxytal AI Lab that orchestrates the entire SDLC as a configurable, auditable pipeline. Each stage is handled by a specialised AI agent. Every handoff is logged. Every critical checkpoint requires human approval before proceeding.
Not replacing developers. Making the boring 40% of their week disappear.
Built for engineering teams
that care about quality.
Not just fast. Auditable, controllable, and transparent at every step of the way.
Critical checkpoints require real sign-off before the pipeline proceeds. Architecture review, code review, and deployment — never fully automated unless you choose it.
Every decision, handoff, and generated artefact is logged with timestamps, agent version, and token cost. Know exactly what happened and why.
Native integrations with Confluence, Jira, GitHub, Vercel, Slack, SharePoint, and the Anthropic Claude API. Works with your existing tools, not instead of them.
Every agent prompt is customisable per project. Branch naming, test coverage targets, code style, documentation level — all configurable from the UI.
See exactly how many tokens each pipeline run costs. Per-agent breakdown, per-project totals, and trend data so you stay in control of AI spend.
Pipeline failed at code review? Fix the issue and resume from that exact stage — no need to re-run requirements extraction or design from scratch.
Eight agents. One pipeline.
Zero handoff delays.
Click any agent to explore what it does and how it fits into the flow.
Reads Confluence pages and extracts every user story — structured, estimated with Fibonacci points, and pushed directly to Jira. Flags ambiguities before a single line of code is written.
How we built ForgePipeline.
From whiteboard to working pipeline in weeks, using the same agile methodology we apply across all Oxytal products.
Mapped the SDLC coordination problem across real engineering teams. Defined the 8 agent stages and the gate approval model.
Designed the pipeline orchestration engine, credential flow, context passing between agents, and the drag-and-drop canvas UI.
Built all 8 agents with real API integrations — Confluence, Jira, GitHub, Vercel, Slack, SharePoint — and a full audit trail at every step.
Deployed at forgepipeline.ai. Actively onboarding early-access customers and iterating based on real pipeline runs.
Works with the tools
your team already uses.
ForgePipeline connects to your existing ecosystem out of the box — no re-tooling required.
Built on a modern,
production-grade stack.
Proven in production.
Real numbers from real pipeline runs on the live platform.
From a single Confluence page to an open GitHub PR with reviewed, committed code — in one automated pipeline run.
Code Review Agent powered by Claude Opus analyses every PR for SQL injection, auth bypass, hardcoded secrets, and OWASP vulnerabilities.
Every agent decision, token cost, and handoff is logged and traceable. Every critical stage requires explicit human approval before proceeding.
Ready to automate
your SDLC?
ForgePipeline is live and accepting early access requests. Join engineering teams already cutting coordination time by 40%.