Practical guides for using Aigon in different modes and workflows.
Drive Mode
Work on a feature solo — one agent, one branch.
Fleet Mode
Multiple agents compete on the same feature in parallel worktrees.
Research Workflow
Multi-agent research with evaluation and feature creation.
Dashboard
The centralised command centre for all repos and agents.
Autonomous Mode
Hands-off feature runs with AutoConductor: implement, review or evaluate, then close.
Feature Sets in Autonomous Mode
How set-level autonomy sequences features, tracks progress, and resumes after failure.
Tutorial: Brewboard Feature Set
A concrete Brewboard walkthrough for tagging a set and running it autonomously.
Scheduling Features
Defer autonomous feature or research starts to an explicit wall time (ISO 8601 with timezone).
Recurring Features
Cadence-driven batches from recurring templates — weekly maintenance, benchmarks, and similar rolls.
Insights
Quality metrics, token analytics, trends, and AI coaching on top of Reports data.
Usage limits (Pipeline)
Agent quota bars, status dot, collapse, and how polling works.
Agent Quota Awareness
Use live quota context to decide which agent should take the implementation work.
Agent Failover
Hand work to the next agent in the chain when the running one hits its quota.
Local Models
Run local LLMs alongside cloud agents using Ollama, LM Studio, or any OpenAI-compatible endpoint.
Aigon Sync
Private Git vault for portable .aigon state — push/pull across machines (CLI + dashboard Settings).
Migration and deprecated guides
- Feedback workflow — legacy feedback commands are no-ops; capture new customer input as research with
origin: customer-feedback.