Released on February 4, 2026, prompt2json v0.3.0 marks a major milestone in the project’s evolution. Previously a Gemini-centric tool, prompt2json has now expanded into a versatile, multi-provider utility, allowing users to bring the power of deterministic, schema-validated JSON to a wider array of Large Language Models.

What’s new

The headline feature of v0.3.0 is the introduction of Multi-Provider Support. By introducing the --provider flag, the tool now supports both Google’s Gemini and OpenAI-compatible Chat Completions APIs.

OpenAI and Local LLM Integration Users can now leverage OpenAI’s state-of-the-art models or run local LLMs via servers like Ollama. This integration brings full support for structured outputs via json_schema, ensuring that the LLM adheres strictly to your defined output shape. For those using OpenAI, a new --strict-schema flag enables “Strict” mode for even more reliable validation.

Universal Control To make the tool more flexible across different environments, v0.3.0 introduces universal API overrides. The --url flag allows you to point the tool at any compatible endpoint, and the --api-key flag provides a consistent way to handle authentication regardless of the provider you are using.

Enhanced Gemini Experience While expanding its horizons, prompt2json continues to support Gemini’s unique capabilities, including the ability to attach images and PDFs to prompts. Additionally, Gemini users can now use the universal --api-key flag as an alternative to Application Default Credentials.

Why it matters

For developers and automation engineers, the value of prompt2json lies in its ability to turn non-deterministic LLMs into reliable interfaces. By expanding provider support, this reliability is no longer locked to a single ecosystem.

Whether you need the massive context windows of Gemini, the widespread availability of OpenAI, or the privacy and cost-efficiency of a local Ollama instance, prompt2json provides a unified CLI experience. You can now swap models and providers while keeping your JSON schemas and shell pipelines exactly the same.

Getting Started and Upgrading

If you are upgrading from a previous version, please note that the --provider flag is now mandatory. To keep your existing workflows running, simply add --provider gemini to your commands.

To try out the new OpenAI support:

  1. Add --provider openai.
  2. Specify your model (e.g., --model gpt-4o).
  3. Provide your key via --api-key or the OPENAI_API_KEY environment variable.

For local LLMs, use the --url flag to specify your local server address.

This post was AI-generated using the model unsloth/gemma-4-31B-it-GGUF:UD-Q5_K_XL. Based on the release of prompt2json v0.3.0 on February 4, 2026. Generated on April 13, 2026. Author: release-storyteller