Skip to main content
Ottili Coder

Coder missions

How a plain-language objective becomes a durable, reviewable run record in Ottili Coder.

A mission is the objective you give Coder in plain language. It is the input that becomes a run.

From mission to run

1. You describe the objective — for example, "add pagination to the orders API and cover it with tests."

2. Coder records the mission as a durable run record*, not a disposable chat message.

3. Planning expands the mission into a reviewable task queue.

4. Execution runs the queue in isolated workspaces.

5. The record closes with logs, approvals, status, and the audit trail kept together.

Writing a good mission

  • State the outcome, not just the steps.
  • Mention the repository area or module.
  • Note any constraints (frameworks, tests, approvals).
  • Keep the scope small enough to verify.

Requirement clarification chat

When you open New Build → Chat Intake*, Coder does not just echo your text back. It normalizes your request into a scoped mission* and asks only the questions that actually change the build — for example which repository, authentication, database, language, user roles, deployment target, or acceptance criteria apply. Each question offers sensible defaults as one-click options.

As you answer, the mission card updates: intent, complexity, build mode, and a confidence score. When nothing build-changing is missing, the card flips to "Mission is clear and scoped. Ready to build."* with two choices:

  • Review technical plan* — opens a structured plan (scope, architecture, APIs, UI, auth, tests, risks, acceptance criteria) you can accept before building.
  • Start build* — dispatches the run directly from the captured goal.

Your answers persist across turns, so revising an earlier answer simply re-normalizes the mission. The conversation always ends in a plan-ready requirements contract, not an endless chat.

Technical contract, endpoint reference, examples, migration and troubleshooting: [Requirement clarification chat](/docs/coder-requirement-clarification-chat).

Mission vs chat

Because a mission is a run record, you can return to it later, see what changed, who approved what, and reuse the queue. It is built for review, not for a one-off conversation.

Related

  • See how missions expand into a [task queue](/docs/coder-task-queues).
  • Choose a [mode](/docs/coder-modes) for how Coder should execute.

Was this article helpful?