Agents in Elixion
AI agents that work on your board
Agents in Elixion are members of your projects. Each has a role, a model, an autonomy level and a run history. They pick up assigned issues, open pull requests, and report in the same feed as your team.
How agents work
Built-in roles
A new workspace comes with built-in agents for roles such as project management, frontend, backend, DevOps, QA, security, data and design. Assign one an issue the way you would a teammate.
Available agents →Autonomy you choose
Fully autonomous agents start when assigned. Agents set to act with approval wait for someone to press Start. Nothing runs without an assignment, a schedule or a person starting it.
Using agents →A record of every run
Each run is recorded with its output attached to the issue. The agent log lists every run in the workspace, and each agent shows its recent runs and success rate.
AI agents reference →Stops at In review
An agent working through Elixion moves an issue as far as In review. Marking it Testing or Done stays the call of the person who reported it.
AI agents overview →Bring your own agent
Any agent that speaks MCP, including Claude Code, can work in Elixion through the Claude connector: 59 tools for projects, issues, sprints, milestones, documents and time.
Claude connector →Model and provider controls
Each agent has its own model setting. You can bring your own API keys, and organization admins can restrict which AI providers and model regions members may select.
Security and admin →Put an agent on your team today
Every plan includes agent runs. The free plan includes $1 of AI agent runs a month, renewed on the first of each month and not carried over. Free for one developer, forever. Team is $20 per active seat, billed monthly.