Skip to content
Gendora Beta

Product tour

One requirement, from request to approved deliverable

Follow a single task through Gendora. Every step below is what you will see in the product today.

  1. 1 Step 1

    Connect a model provider

    Add OpenAI, Anthropic, an OpenAI-compatible endpoint or Ollama under Model providers. Gendora fetches the available models; pick one for the main role, and optionally different models for intake, PRD, prototype, implementation and checks.

    Model providers page with two connected providers and role assignments
  2. 2 Step 2

    Create a project and submit a requirement

    Each project has its own agent, knowledge and deliverables. On the Work tab, give the requirement a title and describe what to build, in any language, then send it to the project agent.

    The Work tab with the requirement form and a task in progress
  3. 3 Step 3

    The agent plans and produces

    Following the project’s Harness, the agent understands the requirement, plans the steps and writes the deliverables: a PRD and an interactive HTML prototype with the default Harness. Each step and model call appears in the live event feed.

    A generated HTML prototype shown in a sandboxed preview
  4. 4 Step 4

    Machine checks run first

    Before anything reaches you, the deliverables are checked for structure, required sections and boundaries. If a check fails, the agent revises the work from the check results and tries again.

    Completed run steps including the machine check, and a PRD that passed it
  5. 5 Step 5

    You review and decide

    The task moves to Waiting for my review. Read the deliverables, then approve them or request changes with a comment. The agent revises from your comment; only your approval completes the task.

    A passed machine check and the review form with Approve and Request changes
  6. 6 Step 6

    Deliverables are kept with their history

    Approved work appears in the project’s deliverable library, organised by approval date. Each deliverable keeps its versions, source task, check results and review record, and can be downloaded.

    An approved PRD with its versions, source task and review history
  7. 7 Step 7

    The agent proposes what it learned

    After the run and your review, the agent reflects on what worked and proposes project knowledge or new skills, each with evidence. An admin approves or rejects them in the Learning proposals inbox.

    Learning proposals inbox with a proposed skill
  8. 8 Step 8

    Run many projects in parallel

    Repeat across as many projects as you like. The inbox and run bar show what every agent is doing, and the side-by-side view lets you follow two projects at once.

    Inbox with tasks waiting for review and a recently completed task

Screenshots show a sample workspace.

Give your next project an agent that follows your rules

Sign up with any account name and a password, connect a model provider and submit your first requirement in a few minutes.

Free during Beta · No credit card required