What is OpenTokenMini?
OpenTokenMini is a free, open-source tool for monitoring and reducing AI agent costs. It measures the cost of each agent call, connects that spending to the commits the agent produced, and returns cost information to the agent so it can spend less. It is intended for people working with Claude Code, Codex, open models, OpenAI, or Anthropic, and can be installed through a coding agent.
OpenTokenMini Features:
Per-Call Cost Tracking: Measures the cost of each AI agent call, giving users a per-call view of spending rather than an undifferentiated total.
Commit Cost Attribution: Connects agent-call costs to the commits the agent produced, associating spending with the resulting code changes.
Cost Analysis: Analyzes agent spending as well as measuring it, so users can examine the costs of their AI agent use.
Agent Cost Feedback: Feeds cost information back to the agent, with the stated aim of reducing how much it spends.
Coding Agent Support: Works with Claude Code and Codex, as well as open models, for people using these coding agents and models.
Model Integrations: Offers integrations for OpenAI and Anthropic, alongside support for open models.
Guided Installation: Can be installed through a coding agent that explains each step and asks before turning anything on.
OpenTokenMini Benefits:
Clearer Spending: Per-call visibility helps users understand where agent spending occurs, replacing guesswork with a clearer view of costs.
Code Cost Context: Commit-linked spending helps users judge the costs associated with code their agent produced.
Lower Agent Spend: Cost feedback gives agents information intended to curb spending, helping users avoid relying on guesswork to manage costs.
No Software Fee: Its free, open-source availability lets users try the cost-monitoring tool without paying for the software.
User-Controlled Setup: Step-by-step explanations and confirmation prompts make setup easier to follow while leaving users in control of what gets turned on.
Use Cases:
Claude Code Monitoring: Review individual call costs while using Claude Code to understand how a coding agent's activity affects spending.
Codex Cost Review: Use with Codex to connect coding-agent costs with commits produced during implementation work.
API Agent Tracking: Monitor agent costs in workflows that use the available OpenAI or Anthropic integrations.
Open Model Workflows: Add cost monitoring to a workflow that uses open models, instead of relying only on hosted model integrations.
Guided Tool Setup: Install through a coding agent when configuring cost monitoring and prefer a setup process that explains steps and requests confirmation.

