Forum runs agent teams with budgets, routing, and restartable records
Forum is a Python for running teams of AI agents. It turns a plain request into smaller tasks, works out which tasks depend on others, and runs independent tasks in parallel. It is not tied to one , so it can use local commands, s, or the Anthropic API.
It supports , where easier work can go to cheaper models first and harder or failed work can move to stronger models. Each run can have limits for model calls and elapsed time, and can estimate, trim, or skip input material to stay within approximate token limits. If a run crashes, checkpoints let it continue from the last completed wave instead of starting over.
Forum also records routing choices, tasks, outputs, and judgments in a ledger that can be checked and replayed, with a hash chain to detect tampering. The same engine can be exposed through HTTP and MCP, has no , and runs on Python 3.11 or newer.
Key points
- Forum breaks a request into agent tasks and can run independent tasks in parallel.
- lets easier work use cheaper models before escalating to stronger ones.
- Run budgets limit model calls and elapsed time, while help manage token use.
- Checkpoints let crashed runs resume without repeating completed task waves.
- A ledger with a hash chain makes past runs replayable and easier to audit.