Skip to content

meetflowstate/flowstate-telemetry

Repository files navigation

flowstate-telemetry

Capture and attribute AI coding tool spend across your engineering organization. flowstate-telemetry configures developer machines to emit OpenTelemetry data from AI tools to Flowstate, giving you project-level cost attribution, usage pattern analysis, and actionable insights to understand and control your AI cloud bills.

One command. Every tool. Full visibility.

Why Flowstate?

AI tool costs are growing fast — Copilot seats, Claude API usage, autonomous agents — but most organizations have zero visibility into which teams or projects drive that spend. Unattributed budgets are the first to get cut.

Flowstate turns AI spend into an attributable, project-level line item:

  • Cost attribution — Break down spend by team, project, and tool across GitHub Copilot, Claude, Cursor, Windsurf, and more
  • Usage pattern analysis — Understand how your engineers use AI tools, identify adoption gaps, and spot cost anomalies
  • Actionable recommendations — Get data-driven suggestions to optimize license allocation and reduce waste
  • Token-level granularity — Track input/output token usage by tool and project for precise cost modelling

Platform Support

Platform Status Install Method
macOS (ARM/x64) Fully supported Homebrew, binary
Linux (x64) Fully supported Binary
Windows (x64) Not yet supported

Supported Tools

Tool macOS Linux Config Method
Claude Code Yes Yes ~/.claude/settings.json env block
Copilot Chat Yes Yes VS Code settings.json
Gemini CLI Yes Yes ~/.gemini/settings.json
Codex CLI Yes Yes ~/.codex/config.toml
Qwen Code Yes Yes ~/.qwen/settings.json
Cursor Yes Yes Hook scripts + hooks.json
Windsurf Yes Yes Hook scripts + hooks.json
Aider Yes Yes ~/.aider.conf.yml
OpenCode Yes Yes ~/.config/opencode/opencode.json

Install

Homebrew (macOS)

brew tap meetflowstate/tap
brew install flowstate-telemetry

Binary (macOS / Linux)

Download the latest release from the releases page and add it to your PATH.

Build from source

make build
make install

Quick Start

# Set your Flowstate API key (get one at https://app.flowstate.inc)
export FLOWSTATE_OTLP_KEY="your-key"

# Auto-detect and configure all installed AI tools
flowstate-telemetry install --all

# Verify everything is working
flowstate-telemetry verify

Usage

Configure all detected tools

flowstate-telemetry install --all

With prompt capture

flowstate-telemetry install --all --prompts

Configure a single tool

flowstate-telemetry install --tool claude-code

Preview changes (dry run)

flowstate-telemetry install --all --dry-run

Check status

flowstate-telemetry status

Remove configuration

flowstate-telemetry remove --all

Fleet Deployment (MDM)

For enterprise rollouts, generate deployment artefacts for your MDM platform:

flowstate-telemetry generate-mdm --format jamf      # Jamf Pro
flowstate-telemetry generate-mdm --format intune     # Microsoft Intune
flowstate-telemetry generate-mdm --format ansible    # Ansible
flowstate-telemetry generate-mdm --format puppet     # Puppet

Environment Variables

Variable Description
FLOWSTATE_OTLP_ENDPOINT OTel collector endpoint (default: https://otel.flowstate.inc)
FLOWSTATE_OTLP_KEY Flowstate API key for authentication
FLOWSTATE_TEAM_ID Team identifier for resource attributes
FLOWSTATE_EMAIL User email for resource attributes

Learn More

License

MIT

About

Flowstate's telemetry CLI

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors