AI infrastructure  /  Cybersecurity & engineering

Inference for security workflows

Build with models that fit your workflow.

Connect your applications and evaluation tools to AI models through a familiar OpenAI-compatible chat completions API. Manage workspace keys, monthly spend budgets, and metered usage from one place.

Request early access →
Headless CLI or local developer token sign-in
Verified server-side against configured JWKS before loading workspace data.

The platform

A clear path from access to integration.

01 / INTEGRATE

Use a familiar API

Connect compatible clients through POST /v1/chat/completions and GET /v1/models. Plain-text chat messages are supported; responses are buffered for safety screening before JSON or SSE frames are emitted.

02 / MANAGE

Control workspace access & spend

Create and revoke workspace-scoped API keys, inspect enabled models, and enforce both prepaid credit balances and aggregate UTC monthly spend budgets in a single-threaded ledger.

03 / REVIEW

Understand your usage

Review uncached input, cached input, and output tokens where reported by the upstream provider. Unreported cache breakdowns and provisional estimates remain explicitly labeled.

Content screening & retention disclosure: To enforce input and output safety policies, DotProd buffers model output until screening completes (even when stream: true is requested) and stores AES-256-GCM envelope-encrypted input/output captures in Cloudflare R2. Captures are accessible only to MFA-authenticated reviewers with tenant-scoped grants and a logged audit reason. Read API & Privacy Terms.

Currently Enabled Gateway Models & Metered Rates

Live model specifications — showing base model lineage, whether weights are abliterated or standard-aligned, and metered rates (USD per 1,000,000 tokens).

View API Reference
Live hosted models, base architecture, abliteration status, and per-million-token pricing
Model ID Based on Abliterated Uncached Input / 1M Cached Input / 1M Output / 1M
Loading live model catalog…

Planning a Kimi K3.1 integration?

Tell us which proposed pricing model fits your workload and we will follow up as capacity opens.

Request early access

Save your new API credentials

Copy the secret API key now. For security, DotProd displays this secret only once and clears it as soon as this dialog closes.


          

Revoke API key?

Clients using this key will immediately receive HTTP 401 Unauthorized. This action cannot be undone.