CLI dev tool to transform raw text into JSON using LLMs. Choose your preferred provider or locally with Ollama.
More CLI tools →74 free dev tools →
CLI tool that converts raw text into structured JSON using LLMs.
- for
- Developers who need to extract JSON from unstructured text via the command line.
- pricing
- free
- license
- MIT
works withOpenAIClaudeMistralOllama
Key features
7 features of JSON THAT
- Schema Guidance — Supports plain-text or JSON Schema to shape the output.
- Multi-LLM Support — Works with OpenAI, Claude, Mistral, and local Ollama models.
- Provider Configuration — Store multiple provider configs and set a default provider.
- Model Selection — Specify model name per provider via flags or env vars.
- Real-time Streaming — Option to stream LLM output as it is generated.
- Line-by-Line Processing — Convert each line of a file or stream into separate JSON objects.
- Multi-line Formats — Handles CSV/JSONL inputs, emitting one JSON record per line.
Use cases
- Transform log lines into structured JSON for downstream analysis.
- Extract entity data from free-form text using a custom schema.
- Convert CSV or JSONL files into normalized JSON objects line by line.
JSON THAT vs alternatives
 JSON THAT |  Claude Code |  Gemini CLI |  Aider |  Ollama | |
|---|---|---|---|---|---|
| Best for | Text-to-JSON conversion with LLMs | AI-assisted coding | General AI terminal agent | AI pair programming | Local LLM hosting |
| Pricing | Free | Subscription | Free | Free | Free |
| DevHunt upvotes | 0 | 0 | 0 | 0 | 0 |
| Launched | Jan 2025 | — | — | — | — |
- JSON THAT vs Claude Code: Focuses on coding assistance in terminal/IDE rather than generic text-to-JSON conversion.
- JSON THAT vs Gemini CLI: Provides a general AI agent for terminal tasks, not specialized for JSON extraction.
- JSON THAT vs Aider: Pairs with developers for code generation, not for arbitrary text parsing to JSON.
- JSON THAT vs Ollama: Runs local LLMs but does not include CLI conversion utilities or schema support.
JSON THAT FAQ
How do I set up my LLM provider?+
Run `jt --setup` and follow the prompts to enter API keys or local Ollama settings.
Can I use a custom JSON schema?+
Yes, pass a schema file with `--schema` using either plain-text or JSON Schema format.
Is there support for streaming output?+
Add the `--stream` flag to see the JSON being generated in real time.
How do I process each line of a file separately?+
Use the `--line` flag; each input line is emitted as an individual JSON object.
Can I choose a different provider per command?+
Specify a provider with `--provider <name>` to override the default for that run.
Summarized by DevHunt from jsonthat.com · Oct 4, 2026. Details may change; check the official site.
.svg?width=64)
