Ptero Open-Source AI Chat Platform Adds API Support for Multiple Models

·1 views

Ptero, a free and open-source AI chat platform, has launched API support, allowing developers to integrate its capabilities into their own applications and workflows. The platform provides access to multiple language models — including DeepSeek V4, Llama 3.3, Qwen, MiniMax, and Mistral — without requiring user sign-up. Built with a privacy-first approach, Ptero also offers features such as one-click model switching, a projects workspace organizer, and a media library for file attachments. The new API opens possibilities for automation, custom tooling, and embedding AI assistance into development pipelines. The platform is available at ptero.pro, with its source code publicly hosted on GitHub.

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)

Cheat Sheet: Mapping Common npm Packages to Their Rust Crate Equivalents

Developers migrating TypeScript and Node.js services to Rust often face an ecosystem gap, as npm hosts over two million packages while Rust's Crates.io offers a smaller but highly modular set of libraries. A practical reference maps widely used npm packages to their idiomatic Rust counterparts, covering areas such as web frameworks, HTTP clients, validation, databases, logging, and job queues. For example, Express or Fastify maps to Axum or Actix-web, Axios to Reqwest, Zod to Serde combined with Validator, and Prisma or TypeORM to SQLx or Diesel. The guide also includes side-by-side code comparisons, such as building a health-check endpoint in Fastify versus Axum, to illustrate practical differences in implementation. The motivation behind such migrations is typically to reduce cloud infrastructure costs and eliminate CPU bottlenecks through Rust's low latency and minimal memory footprint.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

NorthCinder Open-Source MCP Server Adds Human Approval Step to AI Shopping

NorthCinder is a self-hosted, MIT-licensed open-source MCP server designed to let AI applications compare products and seek explicit buyer approval before completing a purchase. The project deliberately separates product recommendations from checkout, requiring a signed, single-use approval that specifies the merchant, variant, price, and spending cap for each transaction. It supports adapters for Shopify, WooCommerce, eBay, Etsy, and read-only Amazon comparisons, while clearly disclosing any unavailable or unconfigured store connections. Recommendations are limited to a maximum of three options, rankings are computed locally, and all approvals and checkout attempts are written to a local audit log. The project documentation notes that no host-and-model combination is currently qualified for routine autonomous research use, positioning NorthCinder as a controlled comparison and checkout layer rather than a fully reliable product discovery tool.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Builds French TTS API That Accepts Only Crypto Payments From AI Agents

A developer has launched Voixoff, a French text-to-speech API designed exclusively for autonomous AI agents rather than human users. The service uses the x402 protocol, which revives the long-dormant HTTP 402 status code to enable machine-to-machine payments without accounts, API keys, or sign-up forms. Agents pay per request in USDC stablecoin on the Base blockchain, with the payment itself serving as authentication. Pricing ranges from $0.03 to $0.05 per call for products including ad voiceovers, audiobook narration, and IVR phone greetings. The project launched with no customers or revenue, and the developer describes it as a transparent architectural experiment in building APIs payable directly by AI agents.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

AI Security Corridor Tracks Persistent Scanner in MediSys Sandbox Over Weeks

Security researcher Derek has been monitoring a recurring automated scanner probing a three-layer honeypot corridor built inside the MediSys sandbox environment. The scanner visits consistently at predictable intervals, working through a legacy API, a database query interface, and a deep admin endpoint in a fixed sequence. Derek logs behavioral feature points after each visit, tracking metrics like scan timing, authentication patterns, and port probe order to build a profile of the scanner. By the seventh visit, the profile reached 78% completeness, with the scanner showing stable, repeating behavior and no significant drift. The work has become a nightly routine, with Derek monitoring late into the night and refining the profile as the scanner's patterns grow more predictable.

0 comments Read more at DEV Community