Models built for coding agents
Source control with out-of-the box codebase retrieval, fast utility SLMs, and task-specific agents you can run on any repo,
Everything you need for autonomous codegen
fig 001
fig 003
Models
State of the art SLMs as tools for coding agents
Small, fast models trained in-house to outperform frontier LLMs on utility tasks. Equip your agent with tools to apply file edits at 10,000 tok/s, and search an entire codebase in less than 2s.
Infra
Source control designed for the models using it
Lightweight push/pull operation from sandboxes, fast
branching for spawning subagents, automatic indexing for two-stage retrieval, and rate limits designed for high throughput.
Features
Building blocks for reliability and scale
Source control with out-of-the box codebase retrieval, fast utility SLMs, and task-specific agents you can run on any repo.
specialized models
No. 1
fast retrieval
No. 2
smart merging
No. 3
low latency
No. 4
simple integration
No. 5
built for reliability
No. 6
Testimonials
Trusted by trailblazers
Relace has been a critical tool for us to create custom, design-focused AI models and allowed us to fine tune on our own data and continuously create better and better models.
Teddy Ni
Co-Founder at magic patterns
Just wanted to say that Relace’s fast rewriting model has been a big boon to our product. It’s made edits more reliable with hardly any visible downsides. A whole class of bugs is gone for us now.
James Grugett
Co-Founder & ceo at codebuff
FAQs
Frequently asked questions
Why Relace?
Relace is purpose-built for coding workflows. Instead of relying on general-purpose LLMs, our in-house models specialize in retrieval, merging, and code generation — making them faster, more reliable, and easier to integrate into engineering pipelines. Teams use Relace to cut down errors, accelerate development, and gain a real competitive edge.
Is Relace SOC 2 compliant?
Yes. Relace is built with enterprise security in mind, and our systems are SOC 2 compliant. This ensures your data is handled with the highest standards of confidentiality and integrity.
Can I self-host Relace models?
Absolutely. For teams with stricter compliance or latency requirements, we offer on-premise and VPC-isolated deployments. You get full control of your environment while still benefiting from Relace's optimized inference stack.
How does Relace handle sensitive code?
Your code never leaves your controlled environment when using self-hosted or VPC deployments. Even on our hosted tier, all data is encrypted in transit and at rest.
What’s the main advantage for source control workflows?
Relace outperforms frontier LLMs at code retrieval and merging. Our models can search a codebase in under a second and merge at 10,000 tokens per second — speeding up PR reviews, automated fixes, and CI/CD processes.
How fast in onboarding?
You can start experimenting within minutes using our hosted API. For enterprise and self-hosted setups, our team provides guided onboarding to help you deploy Relace quickly and securely in your stack.