🎉 Seed RoundWe've raised 2.5M from Swift, YC, & Chapter One. Read more

📚 Case StudyLastro x Metal LLM chatbot

The Retrieval Engine for LLMs

Improve the performance of LLMs on your data with automated processing, enhancement, and retrieval pipelines.

Backed by

Y-Combinator Logo

“Thanks to Metal, our ideas for AI powered applications are now becoming a reality. The platform is easy to use, reliable and fast as hell.”

Matthew Moroz
Head of AI Products & Partnerships, Publicis North America

Our customers


Powerful AI Experiences

Make your AI strategy a reality

Our platform provides everything your team needs to get your LLM application to production.

Versatile Chatbots

Deploy chatbots fully customized with your company's data

Powered by Metal

Faster Search

Search for anything with natural language

Retrieval success
Hallucination detected
Search trends

Monitoring & Alerts

Application and end user observability

Enterprise Security

Secure cloud or private network deployments

Built for Enterprise

Enterprise grade security and support

Actively in SOC 2 audit

SOC 2 Standards

Easily track usage

Analytics & Reporting

Modern auth

API & User Permissions

24/7 support and access to in-house AI experts

White glove support

Modern standards for security


Secure cloud or private network deployments

Run on your own cloud

Learn more

Loved by developers

Easily add AI to your stack

Metal is your production ready, fully-managed, ML Retrieval Platform. Use Metal to find meaning in your unstructured data with embeddings.

Semantic Search to unlock insights and meaning in your unstructured data.

import { Metal } from '@getmetal/metal-sdk';
const metal = new Metal('pk_123', 'ci_123', 'idx_123');
await metal.search({ text: 'hello world'});

Embedding Generation with a single API call. We handle chunking, metadata enrichment, embedding, and indexing.

import { Metal } from '@getmetal/metal-sdk';
const metal = new Metal('pk_123', 'ci_123', 'idx_123');
await metal.index({ text: 'hello world'});

File Uploading to easily chunk, embed, and index multiple file types. We support pdf, xlsx, docx, and more.

import { Metal } from '@getmetal/metal-sdk';
const metal = new Metal('pk_123', 'ci_123', 'idx_123');
await metal.uploadFile('./document.pdf');

Logging & observability to easily track requests, errors, and performance of your retrieval engine.


Read the blog

Check out company updates, tutorials, and AI related learnings.

Introducing our open-source server

🧠 Motörhead: memory for LLMs

Motörhead is a open-source memory server for LLMs. Built in Rust.

Managed LLM retrieval & memory

Get started with a free account