GISELA turns a conversation into requirements, features and scheduled tasks — a full AI agent team plans, codes, reviews and tests them, with your people signing off on every call that matters.
Chat turns ideas into structured requirements — user stories, acceptance criteria, scheduled tasks. The catalogue is your source of truth.
AI engineers pick up tasks, branch, write code and tests against the requirement, and open merge requests — directly in your Git.
Agents hand back reviewed, tested code from day one. In-app deployments and release management follow after the beta.
The beta ships requirements, features, tasks and the full agent team. The in-app features marked later follow on the roadmap.
Every requirement with user story, acceptance criteria and links to tasks and code — one source of truth instead of scattered tickets.
Planner, coder, reviewer and QA are all on board from day one — each owns a role, shows up on the board and is accountable in the activity log.
Don’t like us? Export requirements, features, tasks and all your data as Markdown any AI can understand.
Review agents work from day one — the dedicated in-app review experience lands after the beta.
Semantically searchable code with an in-app code browser — coming after the beta.
Preview, staging, production with pipeline checks and one-click rollback — coming after the beta.
Token spend per agent, cost per completed task and engineering analytics — coming after the beta.
Engineer at heart. Spent years watching good ideas die between ticket and merge — and decided the lifecycle itself is the product.
LinkedIn ↗ — Julian FinklerProduct and platform mind. Convinced that requirements — not prompts — are the language teams should speak with AI.
LinkedIn ↗ — Aljoscha KnaustAt launch, GISELA works with any OpenAI-compatible API — bring your own key and choose the models you trust. Support for Claude and other providers follows shortly after.
EU data residency by default. Your code stays in your Git provider; GISELA indexes it for semantic search and never trains models on your data.
No. Export your requirements, features, tasks and all other data as plain Markdown at any time — a format any AI (and any human) can understand. If GISELA isn’t for you, you leave with everything.
Agents propose — your team disposes. You approve requirements, schedule tasks and review every change an agent hands back before it lands.
Yes. Agents branch, commit and open merge requests directly in your existing GitHub or GitLab repos. No migration, no mirror.
Private beta opens in Q1 2027. It covers requirements, features, tasks and the full agent team — planner, coder, reviewer and QA, all from day one. Separate in-app features like the code review UI, code browser, code map, deployments and insights follow after the beta. Waitlist teams get first access, in order of signup.
Copilots accelerate one developer inside an editor; autonomous agents ship code without much structure around it. GISELA runs the whole lifecycle — requirements, planning, code, review, QA and deployment — as one accountable system your team steers.
Teams on the waitlist get first access and guided onboarding.