Developer Launches Keyless LLM Inference Relay for AI Agents at $0.10 per 1M Tokens

·1 views

A developer has released an OpenAI-compatible LLM inference relay that uses the x402 HTTP payment protocol, allowing autonomous AI agents to access language models without API keys, signups, or credit cards. The service charges $0.10 per million tokens, billed per request in USDC on either Solana or Base mainnet, with gas fees sponsored by the PayAI facilitator. It supports three large-context models — DeepSeek-V4-Flash-0731, Kimi-K2.6, and MiniMax-M2.7 — offering context windows between 180k and 400k tokens. The x402 protocol, revived by Coinbase, works by having the server return a 402 Payment Required response, which the client answers with a signed USDC micropayment before retrying the request. The developer stated the project was built to eliminate the manual API key management burden when running self-directed agents, and is being shared openly for others facing the same friction.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

How to Set Up a Local Kubernetes Cluster on Debian Using Minikube

A step-by-step guide explains how to install Minikube on Debian 12 or 13 (x86-64) using the Docker driver to run a local single-node Kubernetes cluster. The setup requires at least 2 CPUs, 4 GiB of RAM, and 20 GiB of free disk space, along with a non-root user with sudo access. The Docker driver runs the Kubernetes node inside a container, eliminating the need for a separate virtual machine. The guide recommends verifying four layers in order — host capacity, Docker runtime, Minikube profile, and Kubernetes workload — to make troubleshooting easier. The setup is intended for local development and testing, not for production use.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How to Install Docker Engine on Debian Using the Official APT Repository

Docker Engine provides the daemon, CLI, and container runtime required to build and run containers on Debian systems. The recommended installation method uses Docker's official signed APT repository, which allows updates to be managed through the standard package workflow. Supported Debian versions include Bullseye (11), Bookworm (12), and Trixie (13), across amd64, armhf, arm64, and ppc64el architectures. Setup involves three key steps: establishing APT repository trust, installing the engine and plugins, and configuring socket access for the intended user. After installation, running both 'docker version' and 'docker run hello-world' is advised to verify that the CLI and daemon are communicating correctly.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

pidtree 0.6.0 breaks Node process trees on Windows 11 after WMIC removal

Microsoft removed the WMIC command-line tool starting with Windows 11 22H2, silently breaking any Node.js tooling that depends on it for process-tree queries. The widely used pidtree library version 0.6.0 shells out to WMIC on Windows, causing it to fail with an ENOENT error that gets swallowed silently rather than surfaced as a visible crash. Because the failure returns an empty child list instead of an exception, affected applications simply report a process tree of size one, making busy processes appear idle. Version 1.0.0 of pidtree fixes the issue by falling back to PowerShell's Get-CimInstance when WMIC is unavailable, but semver caret ranges pinned to ^0.6.0 will never auto-upgrade to it. Developers are advised to audit their lockfiles for the outdated version and, if upgrading is not possible, replace WMIC calls with a single Get-CimInstance Win32_Process snapshot per polling cycle for better performance.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Why file isolation and cleanup matter more than launch tricks for parallel AI coding agents

Running multiple AI coding agents in parallel on Windows requires careful handling of file isolation, logs, ownership, and cleanup after failed runs — not just a smart launcher setup. Each agent process needs its own set of home directory environment variables, including USERPROFILE, HOMEDRIVE, and HOMEPATH, since Windows does not natively honor the POSIX HOME convention. Shared config files like CLAUDE.md can be linked back to a global copy, but Windows symlinks require elevation or Developer Mode, making hardlinks a common fallback that can silently break 'detach' workflows. Detecting hardlinks requires comparing inode values using Node's bigint stat mode, as standard inode checks return zero on some Windows configurations and produce misleading errors. Git worktrees add another layer of complexity, since forcibly removing a worktree while a process holds an open file handle inside it will fail, requiring cleanup logic to account for locked paths.

0 comments Read more at DEV Community