A drop-in control layer for AI coding tools that blocks secrets, removes wasted context, and routes low-value work to cheaper models.
AI sessions accumulate junk: repeated file reads, duplicate instructions, failed tool output, oversized reasoning blocks, leaked credentials. NeuroRouter strips what doesn't help, so you send less, spend less, and keep sensitive data out of API traffic.
Protect
Secrets never reach the API. Credentials, tokens, and connection strings are intercepted inline. Detected leaks are flagged for rotation — not just blocked, but reported with specific remediation steps so you know exactly what to fix.
Purify
Six filters strip structural noise from every request: stale file reads, thinking blocks, orphaned tool results, failed retries, duplicate system reminders, and oversized content blocks. The API only sees clean context. Real sessions show 40% noise reduction.
Improve
The proxy detects patterns that degrade sessions and surfaces fixes: "Files re-read 5+ times cause 30% waste — try /compact earlier." Not surveillance. Not tracking. Pattern-level feedback that improves every session without changing how you work.
Drop-in replacement for your API endpoint.
ANTHROPIC_BASE_URL=http://localhost:9120 claude
One environment variable. Your agent talks to NeuroRouter. NeuroRouter talks to the API. Everything in between gets cleaned. Your API key is forwarded as-is at the HTTP layer — never parsed, never stored, never logged.
# start the proxy neurorouter --port 9120 # see what would be filtered without sending neurorouter --port 9120 --dry-run
A session sends the same file content 3 times, includes a 120KB thinking block, and repeats system instructions in every request. NeuroRouter strips it before the API call. Less wasted context. Lower cost. Same workflow.
OPS shows how much unnecessary traffic was removed before billing. A session at OPS 75% means 25% of what would have been sent was noise — caught and removed. Track it in your status line or your logs.
Open Source
AGPL v3 — full features, self-hosted
Commercial
$29 / seat / month — pays for itself by reducing token spend
Enterprise
Contact us — volume pricing at 50+ seats