Scores that mean what they say.
teleon.dev/docs/confidence calibrates confidence, strength and probability on separate 0 to 100 scales, and validates proof certificates so a checker can confirm a result quickly.
After calibration, stated confidence tracks how often answers are right.
What it does
Calibrated scales
A score of 80 is right about 80 percent of the time.
Separate scales
Confidence, strength, probability and likelihood stay distinct.
Proofs checked fast
Certificates from formal tools validate in code.
The contract
A score vector names each scale, its value and the calibration it used.
teleon_epistemic_score_vector_0_100/v1task_proof_certificate/v1
Endpoints
| Method | Path | Access | What it does | Status |
|---|---|---|---|---|
POST | /v1/scores/calibrate | RING-2 | Calibrate a raw score | Listed |
POST | /v1/proofs/validate | RING-2 | Validate a proof certificate | Listed |
Listed paths appear in the service’s published files today. In the one-host plan every path here is served under api.teleon.dev/v1/confidence.
How it connects
For agents
Give your agent the read-only guide first. It reports back before any write.
Words on this page
- PAVA
- Pool adjacent violators: the standard way to turn raw scores into calibrated probabilities that only go up.
- Proof certificate
- A small file that lets a checker confirm a result is correct far faster than redoing the work.