DEMO / RECORDING
Overview & Result
When an agent is connected to dozens of MCP servers, feeding all tool schemas into the LLM context causes tool hallucinations and massive token overhead. Jev Tool Router picks only the relevant tools for the prompt.
How Jev fits in the loop
- Inspect full registered MCP tool registry.
- Jev scores and filters down to top 1-3 tool schemas matching user task.
- Codex receives only the relevant tools for execution.
How to reproduce
- Inspect the source implementation at https://github.com/jackbarunz/jev-tool-router.
- Configure local Jev API credentials and register required tool/model bindings.
- Execute test prompts and observe decision latency and routing outputs.
Why this build matters
Solves the notorious 'tool context explosion' problem in agent development.
Limitations
- Metrics are author-reported from the initial release unless independently verified.
- Requires access to the respective agent framework or runtime environment.
- Generative model execution remains external to the Jev decision step.