At Formata, we're helping to build great customer relationships through data, providing Customer Success teams with the knowledge they need to grow revenue, retain customers, and improve their products.

Stof is a data format that unifies structure, validation, and behavior into one coherent, inspectable, portable artifact. We designed Stof to help us create complex data models with AI at the core, required for next-gen customer knowledge management.

As a smart, declarative runtime for data workflows, Stof offers code that acts like data, and data that can think. Working seamlessly with other data formats (import and export), Stof is a low-friction glue-layer between systems and is a real game-changer for data meshes, integration, LLM/AI workflows, smart configs, API definitions, self-describing datasets, data models, pipelines, templates, and data unification.

Modern software (especially AI/ML, infra, cloud, CI/CD, and workflows) increasingly relies on structured data that needs to be:

• Human-readable
• Machine-validatable
• Extendable with logic
• Executable safely
• Translatable between formats
• Transferrable between systems
• Versioned and inspectable

But the tools we have for this are primitive and fragmented:

• JSON/YAML/TOML carry structure, but rely on other tools for behavior, units, schemas, or validations.
• External tools create complexity between systems and often require configuration themselves.
• Configs drift and break across environments.
• Runtime logic and validation are scattered across codebases, devops scripts, and data definitions.

Stof is the next iteration of structured data, bridging the gap between static data and logic to provide a foundation for data in modern, distributed, AI-native computing.

We believe Stof will have a large impact in a few key areas:

• Cloud Infrastructure, DevOps, CI/CD - unification of logic, config, and validation will reduce bugs, drift, and misconfigurations.
• AI/LLM/ML Workflows - pre-processing, pipelines, hyperparameters, and data sources. Also, multi-environment deployment, versioning, and validations.
• Data Engineering - unification, transforms, tagging, schema validation, and pipelines.
• Enterprise Software - products with configs, rules, templates, user-safe scripting, plugins, versioning, & validation.
• API Gateways & Integrations - schemas, transformations, versioned interfaces, & custom behaviors.
• FinTech & Regulation - correctness, validation, access control, provenance, safe logic, change auditing, and compliance.
• IoT & Edge - constraints (units, types), offline contexts, portability (networks), safety, & customizations.
• Infrastructure-as-Data - templates (e.g. generating manifests, docs, policies, configs).
• Software Configuration - per environment variants, validation, upgrades, feature flags/toggles, etc.
• Scientific Computing - data with units, types, and validation for research & presentation.

Stof is currently in early development, we welcome feedback and contributions. The core is stable for experimentation and is actively being used in production at Formata.

Our community is early, but very welcoming and growing quickly. Join our Discord to contribute to the latest discussions and roadmap. Check out our docs for information on how to get started with Stof.

Please star the Stof repository for future development & consider sponsoring the project if you're finding value.

Email us for any general matters, if you'd like to get more deeply involved, or would like us to help you with your project.