intelligence api key
authenticate local kangarouter instances to fetch calibrated priors and earn contributor credits
frictionless connect options:
1. terminal 1-liner
npm / npx
npx kangarouter connect
2. python package
pypi
pip install kangarouter
🤖
tell cursor / claude code / copilot / aider:
configure kangarouter in this repository to coordinate multi-agent routing. run
npx kangarouter connect in the terminal to securely store credentials in .env.local without exposing secrets in chat context.
credit balance
active100
available prior query credits
starter balance:100 credits
earned from telemetry:+0 credits
sparse cohort query:FREE
verified run reward:+5 credits
mature query cost:1 credit
route prior query sandbox
query empirical bayesian posterior distributions for any task shape in real time
bounded_planning|LOW|S|BOUNDED|READ_ONLY
sparse_contributor_free (0 credits)
| candidate (route / model) | wilson 95% interval | point est. | expected p95 | differential privacy | status |
|---|---|---|---|---|---|
| querying edge priors... | |||||
privacy, sovereignty & erasure protocol
gdpr compliantzero-code guarantee:
kangarouter operates strictly over anonymized execution metadata digests. prompts, code diffs, file trees, and credentials are hard-rejected by closed schema validators.
telemetry revocation:
you may purge your contributed receipts at any time via POST /v1/telemetry-erasure using your authenticated api key.