Felan Code
An open-source coding agent that runs locally with your own model provider credentials.
npm install --global @felan-ai/felan && felanEfficiency Only Counts When the Task Succeeds
Compare model cost only for runs that meet the task's quality requirements.
Keep context focused
Load project instructions where they apply. Keep research and tool output bounded.
Make progress inspectable
Track subagents, dependent tasks, and decisions as the work progresses.
Measure verified work
Measure cost per verified task, not token count alone.
Cost Reductions With Quality Gates
In seven controlled, extension-specific comparisons, candidate configurations used 41.9% less aggregate model cost than their baselines. Every candidate met its configured quality gate.
Each extension was measured separately against its disabled baseline, not as one all-enabled configuration. Results depend on the evaluated tasks and setup.
Review methodology and resultsAggregate model cost reduction
41.9%
Across seven separately measured extensions. Quality gates passed for every candidate configuration.
One Agent for Local and Team Work
Felan Code runs locally with provider-owned credentials; no Felan account is required. Felan's cloud platform uses the same agent for team workflows.
Runs With Your User Permissions
The local agent is a host application, not a sandbox. It uses your filesystem and process permissions. Use an isolated host for untrusted projects or commands.
Read the runtime and security guideKeep Your Attention for the Hard Problems
Start with one repo and one recurring task.