TokenEconomizer

TokenEconomizerLink to this section

TokenEconomizer is an agent framework - Experience-Driven Black-Box Cost Control for Coding Agents.

OverviewLink to this section

Aaron J. Li 1 , Weijia Xu 2 , Chandan Singh 2 , Zixiao Chen 3 , Jeevana Priya Inala 2 , Charlotte Siska 3 ,

Our measurement on SWE-Bench tasks: 18.4–30.9% lower mean cost across five agent–model combinations, cheaper in all 30 repository × combination cells, and Pass@1 up in 4 of 5 . Results

TokenEconomizer makes coding agents cheaper to run in the repositories you work in often, without changing the agent itself. It learns from the sessions an agent has already run in a repository, and uses that experience to cut waste in the sessions that follow. It has three mechanisms:

TokenEconomizer is open-source, written primarily in Python, with 4,788 GitHub stars under the Apache 2.0 license. The latest release is v0.1.0 (2026-10-26).

Key capabilitiesLink to this section

From the project's documentation:

• Micro-skills: remember how the repository works. Sessions keep rediscovering the same
• Tool pruning: stop loading tools that go unused. Every loaded tool's definition is sent with
• Failure-loop intervention (the loop guard): step in when a turn is stuck. When a turn (the
• Work through what the agent already reads. Facts enter the context at session start, and
• Count cost as it is billed. Most tokens in a session are cached re-reads, so cost is
• Protect task success. Facts form only from repeated evidence, tools are removed only after

How it fits a local-AI stackLink to this section

TokenEconomizer runs on your own hardware, so pair it with a model and a GPU sized to your needs. Use the VRAM calculator to pick a model that fits your card, and see what you can run for hardware guidance. Related agent frameworks in the directory:

• "Agent-Reach"
• "agentic-inbox"
• Agno
• "ai-agents-from-scratch"

SourcesLink to this section

• Source code & docs: microsoft/TokenEconomizer

Stats from GitHub, 2026-11-05.