Inferiq
Inferiq is an early-access software platform for coding agents: it provides one API key and OpenAI-compatible gateway for the coding models it serves, with fallback, metering and receipts. It also provides specialist fix agents that turn coding errors into a cause, patch and verification steps.
At a Glance
- Individual developers using AI coding tools
- Teams running coding agents or scripts
- Developers using Cline, OpenCode, Cursor or Claude Code
- Engineering workflows involving npm, pnpm, Next.js, Vite, Docker or Vercel
AI Tools by Inferiq
(1)
Inferiq
OpenAI Compatible Gateway and Fix Agents
Discussions
No discussions yet
Be the first to start a discussion about Inferiq
Products & Services
A hosted OpenAI-compatible gateway at api.inferiq.run/v1 for coding tools. It routes to Inferiq-hosted Qwen3.8-27B, supports streaming and tool calls, retries eligible failures, applies key/model/rate and spending controls, charges actual usage and writes receipts.
A fix agent that accepts an error and stack trace plus relevant files and returns the cause, a small patch and how to verify it.
A fix agent for npm, pnpm, Next.js, Vite, Docker and Vercel install, build and deployment failures; it reads build logs, manifests, lockfiles and configuration and returns a patch and passing command.
Remote MCP service at https://api.inferiq.run/v1/mcp that lets Claude Code, Cursor and other MCP clients quote and run fixes and check balance from the editor.
Market Position
Inferiq positions itself as a coding-focused model gateway plus debugging specialists, rather than only a general multi-model router. Its differentiators are a single key, hosted Qwen3.8-27B, fallback/retry, per-request receipts, prepaid wallet balance and fix agents that return verified patches without running commands locally. It overlaps with general model gateways such as OpenRouter, which also offers one endpoint, many models and automatic fallbacks, but Inferiq is narrower and adds the coding-error-to-patch workflow.
Founding Story
Inferiq says it was built to solve two problems for developers using AI coding agents: fragmented provider keys, accounts, billing and error formats; and agents getting stuck in repeated patching loops. Its initial vision was to build the gateway and shared routing/metering/receipt infrastructure first, then put specialist fix agents on top of it.
Business Model
Revenue Model
Pay per call from a prepaid USD balance. Model requests are charged by actual input/output tokens; fix agents are quoted before execution, held against the balance, and charged for actual work up to the approved maximum. Inferiq states there is no subscription or monthly minimum; deposits incur a 3% fee.
Pricing Tiers
The same model is exposed through inferiq/fast, inferiq/balanced and inferiq/best; deeper reasoning can produce more output tokens.
Qwen3.8-27B with light reasoning.
Qwen3.8-27B with medium reasoning.
Qwen3.8-27B with full reasoning.
Qwen3.8-27B with light reasoning.
Qwen3.8-27B with medium reasoning.
Qwen3.8-27B with full reasoning.
Target Markets
- Individual developers using AI coding tools
- Teams running coding agents or scripts
- Developers using Cline, OpenCode, Cursor or Claude Code
- Engineering workflows involving npm, pnpm, Next.js, Vite, Docker or Vercel
- Daily coding with AI tools while using one model key and tracking session cost
- Running custom coding agents or scripts through one endpoint
- Keeping an AI coding session running through provider rate limits or outages
- Getting a specialist diagnosis and patch when an agent is stuck in a repeated debugging loop
- Calling fixes from Claude Code, Cursor or another MCP-capable editor
- Predictable developer spend through prepaid balance and per-key limits