PredictLM API

Tabular prediction over HTTPS.

Zero-training regression and classification in a single call — with calibrated uncertainty, feature attributions, and an optional plain-language explanation. Self-serve: create an account and get a key in seconds.

Create account Sign in Open dashboard

Free while in research preview — no card required.

Quickstart

Three calls to your first prediction.

bash

curl -s https://predictlm-api.zerooneresearch.ai/auth/register \
  -H 'content-type: application/json' \
  -d '{"email":"you@company.com","password":"a-strong-password"}'
# -> { "access_token": "...", "plan": "free" }

Issue an API key (or copy one from the dashboard)

bash

curl -s https://predictlm-api.zerooneresearch.ai/auth/api-keys \
  -H 'authorization: Bearer <access_token>' \
  -H 'content-type: application/json' -d '{"name":"prod"}'
# -> { "api_key": "ml_sk_...", "note": "shown only once" }

Predict

bash

curl -s https://predictlm-api.zerooneresearch.ai/v1/predict \
  -H 'X-API-Key: ml_sk_...' -H 'content-type: application/json' \
  -d '{"X_train":[[5.1,3.5,1.4],[6.2,2.9,4.3]],
       "y_train":[0,1],
       "X_query":[[5.9,3.0,5.1]]}'
# -> { "predictions": [1], "task_type": "classification", ... }

Endpoints

Predict, or predict and explain.

POST /v1/predict

The prediction only. In-context learning in one forward pass — float targets route to regression, integers to classification with calibrated probabilities. Uses 1 unit of your monthly allowance.

POST /v1/explain

Prediction + calibrated uncertainty + ranked feature attributions, and an optional plain-language rationale. Pass "want":["attributions"] to skip the LLM (zero tokens). Uses 5 units.

Authenticate with an X-API-Key header (or a Bearer JWT for the dashboard). All traffic is HTTPS. Calls count against your monthly allowance — watch it live in the dashboard.

Model

Best model, same contract.

The endpoint always runs the current-best checkpoint. When a stronger model clears our public-benchmark and latency gates, it swaps in behind the same v1 contract — your code doesn't change.

Currently serving

PredictLM‑Base

26M · Apache-2.0 · contract v1

Weights on Hugging Face →

2026-06-26PredictLM‑Base — Self-serve launch — /v1/predict and /v1/explain go live behind the v1 contract.

Access

Free while in research preview.

No pricing tiers, no card, no billing. Every account gets a monthly allowance of calls while we run the research preview.

Research preview

€0 — free for now

1,000 units / month per account

A prediction uses 1 unit, an explanation 5

Need a bigger allowance? Tell us what you're building and we'll raise it — free stays free.

Enterprise

Need volume, on-prem, or a custom SLA?

Self-serve covers most teams. For higher throughput, private deployment, or a larger allowance, tell us what you're building and we'll get back to you.