Relace

What it does

Relace provides API access to open-weight inference models, including DeepSeek V4 Pro, DeepSeek V4 Flash, DeepSeek V4.1 Flash, Kimi K3, GLM 5.3, GLM 5.3 Flash, and GLM 5.2. The official site directs customers to sign up for an API key to use the models directly; they are also available through OpenRouter. The pricing page describes per-token billing for input, cached input, and output, with prices subject to change.

Its documentation also describes model and coding-agent tools: Compact compresses long agent traces; Fast Agentic Search searches codebases through a pre-built agent harness; Instant Apply merges LLM-generated code snippets into an existing codebase. The docs further list Relace Repos, described as GitHub for AI agents with built-in agentic search, and specialized agents that act on those repos. The published materials are aimed at developers building or operating coding agents and other inference workloads. They identify an interactive playground and model guides, but do not establish supported client platforms or explain detailed setup requirements beyond account creation.

Start by creating a Relace account, then try the playground or follow a model guide; for direct inference, obtain an API key. The listed models and prices may change, and volume pricing is available by contacting Relace. The evidence does not specify technical compatibility, rate limits, or whether all documented tools are generally available.

What it can do

• API inference with listed open-weight models
• Interactive model playground
• Compress long agent traces with Compact
• Codebase search using a pre-built agent harness
• Apply LLM-generated code snippets to an existing codebase
• Agent-oriented repositories and specialized agents

Works with

• OpenRouter

Before you choose

• Model rates are per million tokens and cover separate input, cached-input, and output meters; the API-time price may differ as rates are repriced.
• The materials do not specify supported client platforms, rate limits, or technical compatibility.
• Volume pricing is available by contacting Relace; the sources do not describe a recurring free tier.

Listed under

• Coding
• Sandboxes and infrastructure
• API
• Coding agents
• Developer tools
• Inference
• LLM infrastructure
• Open-weight models

Get started

The company site directs users to sign up and get an API key; docs also recommend trying the playground.

1. Create a Relace account.
2. Try the interactive playground or consult a model guide.
3. For direct model use, obtain an API key; alternatively, use the models through OpenRouter.

No subscription or minimum. Volume pricing is available by contacting Relace. The API price at request time is the amount charged. These rates cover the named inference models; confirm pricing for specialized coding services separately.

Source: relace.ai/pricing, checked 1 Nov 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 1 Nov 2026. Each source supports part of the overview; availability and prices can change.

• Relace ↗relace.ai
• Introduction - Relace ↗docs.relace.ai
• Pricing - Relace ↗relace.ai

Company website

relace.ai: product, pricing and policy pages

Hacker News

Launch posts and discussion

YC directory mirror

Batch and company name

Human review

Checked by a person

manual_review

Last checked 1 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.