top of page

FTAI — Foundational Traceable AI Interface

Why FTAI?

FTAI brings together the best aspects of existing formats while solving their key limitations. It's designed from the ground up for AI-native workflows, offering human readability without sacrificing machine parseability. Whether you're building autonomous agents, structuring training data, or creating reusable prompt templates, FTAI provides the foundation you need.

  • Deterministic Parsing - Every FTAI document yields identical structure across all systems and platforms

  • Built-in Traceability - Embed rationale, constraints, and decision context directly in your data

  • Multimodal Support - Native handling of images and visual content with @image tags

  • Minimal Syntax - Clean, readable markup without verbose punctuation or nested complexity

  • Seamless Interoperability - Converts cleanly to and from JSON, YAML, and plain text formats. 

  • Feature.                          FTAI        JSON        YAML       Markdown

  • Human Readable.   ⭐⭐⭐  |  ⭐             | ⭐⭐    |    ⭐⭐⭐

  • Machine Parse         ⭐⭐⭐   |  ⭐⭐⭐  |  ⭐⭐   |    ⭐

  • AI-Native                     ⭐⭐⭐   |  ❌            |  ❌        |    ❌

  • Multimodal                 ⭐⭐⭐   |  ❌            |  ❌        |     ⭐

  • Traceability                ⭐⭐⭐   |  ❌            |  ❌        |     ❌

Use Cases & Applications

FTAI serves as the communication backbone for modern AI systems. From multi-agent coordination to knowledge management, FTAI provides structure where it matters while maintaining the flexibility developers need. Organizations use FTAI to standardize their AI workflows, create reproducible experiments, and build systems that scale from prototype to production.

Examples:

  • AI Agent Communication - Standardized format for agent-to-agent information exchange

  • Prompt Engineering - Create versioned, reusable prompts with embedded context

  • RAG Systems - Structure knowledge bases with rich metadata for retrieval

  • Workflow Automation - Define complex AI workflows with clear dependencies

  • Training Data - Format examples with metadata for better model performance

Getting Started

FTAI is designed to be immediately accessible. Install the Python package with pip, explore the examples in our GitHub repository, or try the interactive playground. The specification is fully documented, and our linting tools help ensure your FTAI documents are valid and well-formed. Whether you're building your first AI application or integrating FTAI into existing systems, you'll find the tools and documentation you need.

To install FTAI, use pip:​

Copies the install command. Paste into your terminal. For advanced options, view on GitHub.

 

Or visit our GitHub repository for Swift, JavaScript, and other implementations.

Frequently Asked Questions (FAQ)

Q: What is FTAI?
A: FTAI (Foundational Traceable AI Interface) is an open-source data format specifically designed for human-AI communication. It combines the readability of Markdown with the structure of JSON, while adding AI-specific features like built-in rationale tracking and multimodal support.

Q: How does FTAI compare to JSON?
A: FTAI is more human-readable than JSON while maintaining full machine parseability. It eliminates JSON's verbose syntax and adds AI-native features like embedded rationale and native image support. FTAI documents also convert cleanly to JSON when needed.

Q: Who should use FTAI?
A: FTAI is built for AI developers, researchers, agent architects, and anyone building AI-native applications. If you're working with LLMs, multi-agent systems, or AI workflows, FTAI provides the structured communication layer you need.

Q: Is FTAI production-ready?
A: Yes! FTAI v1.0 was released with full linting tools, parsing libraries for multiple languages, and comprehensive documentation. It's actively maintained by FolkTech AI and used in production systems.

Q: Where can I find examples and documentation?
A: Visit the GitHub repository at github.com/FolkTechAI/ftai-spec for complete documentation, extensive examples, and the full specification. The repo includes test vectors, parser implementations, and real-world usage examples.

Resources &

Support

FTAI is built in the open with an active community of contributors. Whether you're just getting started or building enterprise applications, you'll find the support and resources you need. Star the project on GitHub, join discussions, report issues, or contribute to the specification. For enterprise consulting and priority support, contact FolkTech AI directly.

To learn more about FTAI or even download it to use for your self please visit our GitHub page

bottom of page