TensorZero

TensorZero creates a feedback loop for optimizing LLM applications — turning production data into smarter, faster, and cheaper models.

1. Integrate our model gateway
2. Send metrics or feedback
3. Optimize prompts, models, and inference strategies
4. Watch your LLMs improve over time

It provides a data & learning flywheel for LLMs by unifying:

• Inference: one API for all LLMs, with <1ms P99 overhead
• Observability: inference & feedback → your database
• Optimization: from prompts to fine-tuning and RL (& even 🍓? →)
• Experimentation: built-in A/B testing, routing, fallbacks

Website · Docs · Twitter · Slack · Discord

Quick Start (5min) · Comprehensive Tutorial · Deployment Guide · API Reference · Configuration Reference

Demo

Note

Watch LLMs get better at data extraction in real time with TensorZero!

Dynamic in-context learning (DICL) is a powerful inference-time optimization available out of the box with TensorZero. It enhances LLM performance by automatically incorporating relevant historical examples into the prompt, without the need for model fine-tuning.

LLMs-get-better-at-data-extraction-in-real-time-with-TensorZero.mp4

Overview

1. The TensorZero Gateway is a high-performance model gateway written in Rust 🦀 that provides a unified API interface for all major LLM providers, allowing for seamless cross-platform integration and fallbacks.
2. It handles structured schema-based inference with <1ms P99 latency overhead (see Benchmarks) and built-in observability, experimentation, and inference-time optimizations.
3. It also collects downstream metrics and feedback associated with these inferences, with first-class support for multi-step LLM systems.
4. Everything is stored in a ClickHouse data warehouse that you control for real-time, scalable, and developer-friendly analytics.
5. Over time, TensorZero Recipes leverage this structured dataset to optimize your prompts and models: run pre-built recipes for common workflows like fine-tuning, or create your own with complete flexibility using any language and platform.
6. Finally, the gateway's experimentation features and GitOps orchestration enable you to iterate and deploy with confidence, be it a single LLM or thousands of LLMs.

Our goal is to help engineers build, manage, and optimize the next generation of LLM applications: systems that learn from real-world experience. Read more about our Vision & Roadmap.

Get Started

Start building today. The Quick Start shows it's easy to set up an LLM application with TensorZero. If you want to dive deeper, the Tutorial teaches how to build a simple chatbot, an email copilot, a weather RAG system, and a structured data extraction pipeline.

Questions? Ask us on Slack or Discord.

Using TensorZero at work? Email us at [email protected] to set up a Slack or Teams channel with your team (free).

Work with us. We're hiring in NYC. We'd also welcome open-source contributions!

Examples

We are working on a series of complete runnable examples illustrating TensorZero's data & learning flywheel.

Writing Haikus to Satisfy a Judge with Hidden Preferences

This example fine-tunes GPT-4o Mini to generate haikus tailored to a specific taste. You'll see TensorZero's "data flywheel in a box" in action: better variants leads to better data, and better data leads to better variants. You'll see progress by fine-tuning the LLM multiple times.

Improving Data Extraction (NER) by Fine-Tuning a Llama 3 Model

This example shows that an optimized Llama 3.1 8B model can be trained to outperform GPT-4o on a Named Entity Recognition (NER) task using a small amount of training data, and served by Fireworks at a fraction of the cost and latency.

Improving LLM Chess Ability with Best-of-N Sampling

This example showcases how best-of-N sampling can significantly enhance an LLM's chess-playing abilities by selecting the most promising moves from multiple generated options.

Improving Data Extraction (NER) with Dynamic In-Context Learning

This example demonstrates how Dynamic In-Context Learning (DICL) can enhance Named Entity Recognition (NER) performance by leveraging relevant historical examples to improve data extraction accuracy and consistency without having to fine-tune a model.

Improving Math Reasoning with a Custom Recipe for Automated Prompt Engineering (DSPy)

TensorZero provides a number of pre-built optimization recipes covering common LLM engineering workflows. But you can also easily create your own recipes and workflows! This example shows how to optimize a TensorZero function using an arbitrary tool — here, DSPy.

& many more on the way!