Lispa: cloud coding teams with repository review and model routing

Start with a repository task

Lispa is a cloud workspace where teams of coding agents build, test and review work in a connected repository. The user supplies an outcome, and the platform organizes tasks and coordinates agents around that brief.

The site describes cloud workspaces that keep running when a laptop is closed. Its review surface places proposed changes beside agent review and CI evidence, with the user's team deciding whether to accept the work or request changes.

Route different parts of the work

Lispa supports connected subscriptions, API keys and its own credits. Its model-routing approach uses stronger models for difficult work and lower-cost models for routine steps while retaining the task in the same workspace.

The publisher presents an internal routing benchmark with lower model-call costs, but explicitly excludes Lispa fees and machines and labels the example historical. Treat those figures as evidence about that measured setup, not a forecast for a new project's bill.

The captured starting plan is $20 per month including $20 in usage credits. Teammates can share a credit balance, with per-member limits. Provider subscription limits still apply, while machine and service use can consume credits separately.

Evaluate one inspectable delivery

Franklin analysis: a first task should name the branch, expected result and relevant checks. A bug reproduction with a regression test makes it easier to judge whether the agents fixed the actual cause.

Inspect the diff, test output, model use and recorded charges before expanding to a recurring review or maintenance workflow. The site's examples deliberately stop short of an automatic merge, making delivery and the shipping decision separate steps. Verify those settings for the connection you configure. Lispa's useful proposition is a coordinated task with reviewable evidence; the team's assessment of that evidence remains the point where a proposed change becomes an accepted one.