Documentation
Lineman runs a secondary LLM that compresses the data-heavy parts of your Claude Code session (large file reads, build output, logs), so your model spends its context on reasoning, not raw bytes. Everything below is what you need to install it, keep it up to date, and run it.
Start here
Lineman runs a second, cheaper model over the data-heavy parts of your Claude Code session, so your model spends its context on reasoning rather than raw bytes. Install it in one command, then keep working exactly as you do now.
npx -y @lineman-io/cli@latest setupIntegrations
- n8nBring your AI-coding cost data into any n8n workflow with the Lineman community node.
- ZapierPull spend, savings and cost alerts into any Zap, and drive budgets from an automation.
- GitHubAttribute AI spend to your repositories and branches, and flag repos your team wasn't expected to work in.
- JiraShow each ticket's cumulative AI spend on the Jira issue your team already reads.
- Reporting connectorAsk Claude about your Lineman spend from Claude Desktop or Claude Code.
More from Lineman
- QuickstartThe five-step install walkthrough, with troubleshooting.
- BenchmarksMeasured token savings and output-quality results.
- SecurityHow your code and prompts are handled in transit and at rest.
- FAQShort answers to the questions we hear most.
- PricingPlans, token allowances, and the 14-day free trial.
- DashboardYour usage, savings, plan, and API keys.
Support
Last updated: July 2026