.txt
To make AI programmable and composable by providing structured, reliable outputs that meet the reliability demands of real-world systems.
At a Glance
- AI Engineers
- DevOps Teams
- Enterprise AI teams
- Inference Providers
AI Tools by .txt
(1)
Outlines
LLM Structured Output Library
Discussions
No discussions yet
Be the first to start a discussion about .txt
Latest News
Dottxt raises $11.9 million in funding to tell AI models how to answer
Announcing Our Investment in dottxt: From Back-and-Forth Prompts to Truly Dependable Computers
dottxt closes $3.2M Pre-seed round led by Elaia
Outlines library exceeds 65 million downloads
Products & Services
Popular open-source Python library for structured text generation using constrained decoding.
Cloud platform (api.dottxt.ai) offering high-performance, OpenAI-compatible structured generation endpoints.
Managed and self-hosted versions of constrained decoding libraries (dotjson, dotgrammar, dotlambda) for integration with vLLM, TensorRT-LLM, etc.
Market Position
Differentiates by being model-agnostic and offering higher reliability and speed than native structured output features of providers like OpenAI.
Leadership
Founders
Rémi Louf
Co-founder & CEO. Previously Senior Research Engineer at Normal Computing and Research Engineer at Hugging Face. Background in Quantum Physics, Statistical Physics, and Philosophy of Physics from ENS and Oxford.
Dan Gerlanc
Co-founder & CTO. Previously VP of Engineering at Normal Computing and Head of Data at Ampersand. Former hedge fund quant with 15+ years of experience. Education from Williams College.
Brandon T. Willard
Co-founder & Chief Science Officer. Previously Staff AI Research Engineer at Normal Computing and Data Science Lead at Ampersand. Expert in applied math, statistics, and symbolic computing.
Executive Team
Rémi Louf
CEO
Former Senior Research Engineer at Normal Computing and Hugging Face.
Dan Gerlanc
CTO
Former VP Engineering at Normal Computing and Head of Data at Ampersand.
Board of Directors
Founding Story
Founded by former researchers at Normal Computing who realized that LLM outputs were too unpredictable for production systems. They first created the 'Outlines' library and then launched .txt to commercialize the technology.
Business Model
Revenue Model
Usage-based (pay-per-token) for API and licensing/subscriptions for Enterprise/Self-hosted solutions.
Pricing Tiers
Access to Outlines OSS library.
Pay-per-token for high-performance structured generation via api.dottxt.ai.
Self-hosting, dedicated support, and custom integrations.
Target Markets
- AI Engineers
- DevOps Teams
- Enterprise AI teams
- Inference Providers
- Information extraction
- Classification
- AI Agent workflows
- Data transformation
- Structured data generation
- Nvidia
- Hugging Face
- LMU Klinikum