GitHub for Agents
Repos is a source control product co-optimized with our suite of fast, specialized coding models. It's git compatible, high throughput, and has built-in codebase retrieval.
Relace offers a free tier and hosted API that enables experimentation within minutes. Enterprise and self-hosted setups include guided onboarding, and the simple integration model allows teams to start using specialized models immediately without extensive infrastructure changes.
Relace's in-house SLMs are purpose-built for coding workflows and deliver faster inference than general-purpose LLMs while reducing token consumption. Token-based pricing ($0.05–$3.00 per million tokens depending on model) scales with actual usage, lowering costs for high-throughput agent deployments.
Relace provides specialized SLMs trained for code generation tasks (apply, search, rank, embed) that outperform general-purpose LLMs on utility tasks. The fast-apply model merges file edits at 10,000 tokens/second and semantic search scales to large codebases in under 2 seconds, enabling reliable autonomous codegen at scale.
Relace Repos provides source control designed for agents with lightweight push/pull operations, no rate limits, and automatic indexing for two-stage retrieval. The fast-apply universal code merging model and best-in-class semantic search enable rapid PR reviews and automated fixes without performance bottlenecks.
Relace is SOC 2 compliant and offers self-hosted, on-premise, and VPC-isolated deployment options where code never leaves the controlled environment. Data is encrypted in transit and at rest, providing teams with full control while maintaining optimized inference performance.