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
Repos
Source control designed for agents with light weight push/pull operations and no rate limits.
Code Retrieval
Best in class semantic search that scales to large codebases out-of-the-box.
Fast Apply
Universal code merging model that applies file edits at 10,000 tok/s.
Everything you need for autonomous codegen
Repos
Source control designed for agents with light weight push/pull operations and no rate limits.
Code Retrieval
Best in class semantic search that scales to large codebases out-of-the-box.
Retrieval Recall@k (%)
Fast Apply
Universal code merging model that applies file edits at 10,000 tok/s.
Fastest Model by Provider (tok/s)
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.
PendingPendingSource 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.
Building blocks for reliabilty and scale
Specialized Models
Purpose-built models for code retrieval, merging, and generation that outperform frontier LLMs on autonomous codegen subtasks.
Fast Retrieval
Search codebases quickly with built-in two stage retrieval that enables 10x larger embed/rerank models for more accurate results.
Smart Merging
Our apply model processes 10,000 tokens per second to integrate AI edits with existing code.
Low Latency
Optimized inference stack delivers fast response times for practical use in development workflows.
Simple Integration
Drop-in API for CI/CD pipelines, code review tools, and existing development workflows.
Built for Reliability
Designed for production environments where code correctness and consistent results matter.
Building blocks for reliabilty and scale
Specialized Models
Fast Retrieval
Smart Merging
Low Latency
Simple Integration
Built for Reliability
Trusted by the 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.
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.
Just wanted to say that Relace's fast rewriting model has been a big boon to our product. It's made edits much more reliable with hardly any visible downsides. A whole class of bugs is gone for us now.
Just wanted to say that Relace's fast rewriting model has been a big boon to our product. It's made edits much more reliable with hardly any visible downsides. A whole class of bugs is gone for us now.
Do you guys ever sleep?
Do you guys ever sleep?
This is definetely a lot faster!
This is definetely a lot faster!
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.
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.
Just wanted to say that Relace's fast rewriting model has been a big boon to our product. It's made edits much more reliable with hardly any visible downsides. A whole class of bugs is gone for us now.
Just wanted to say that Relace's fast rewriting model has been a big boon to our product. It's made edits much more reliable with hardly any visible downsides. A whole class of bugs is gone for us now.
Do you guys ever sleep?
Do you guys ever sleep?
This is definetely a lot faster!
This is definetely a lot faster!
Frequenty Asked Questions
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.
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.
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.
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.
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.
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.
Frequently Asked Questions
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.
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.
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.
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.
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.
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.
Get Started in Minutes
Try out our playground, and start with our free tier to test Relace models in your application.
Sign up for free