Independent AI API proxy

Claude API for Cursor

Connect Cursor to supported Claude models through CorvusLLM with a documented endpoint shape, prepaid balance, current public model slugs, and clear service boundaries.

Independent service. Not affiliated with OpenAI, Anthropic, Google, or Z.AI.

Get API Key Open Setup Docs
Prepaid balanceUsage is deducted from your CorvusLLM account balance.
Card, wallet, or crypto checkoutAvailable payment methods are shown before order creation.
No financially backed SLAUse the status page and a small pilot before larger usage.
Direct answer

Use this page when Cursor needs supported Claude access.

CorvusLLM can fit Cursor users who need an independent prepaid access layer for supported Claude models, a clear base URL, current public model slugs, pricing proof, and setup documentation before they send real prompts.

  • Use docs for exact fields and screenshots when the tool changes.
  • Use the model catalog for the current customer-facing slug.
  • Run one small request before repository-wide or production usage.
  • Route private key, order, payment, and balance issues to support.
Setup path

Configure Cursor without guessing values

Open Cursor settings, add a custom OpenAI-compatible provider when your build exposes that option, then paste the CorvusLLM base URL, key, and model slug.

Install requirement

Install Cursor, keep your project secrets out of committed files, and verify that your Cursor build exposes custom provider or custom base URL controls.

Model family fit

Claude models are usually a fit for coding agents, reasoning-heavy writing, analysis, refactors, and assistant workflows that benefit from Claude-family behavior.

Important caveat

Claude-native clients can use Anthropic-style request shapes, while OpenAI-compatible clients should use the OpenAI-style CorvusLLM endpoint and public model slugs.

base_url = "https://base.corvusllm.com/v1" api_key = "YOUR_CORVUSLLM_KEY" model = "claude-haiku-4-5"
Field Use this Why it matters
Where to edit Cursor Settings -> Models or custom provider controls when your build exposes them. Use the tool-owned settings area instead of hardcoding keys in prompts or visible node text.
Base URL https://base.corvusllm.com/v1 Use the OpenAI-compatible route for SDKs, Open WebUI, Cursor-style custom providers, ChatBox, n8n, Windsurf, and similar clients.
API key Your CorvusLLM key Keep it in the tool secret store, credentials area, or server environment variables.
Model slug claude-haiku-4-5 or claude-opus-4-5 Use public Claude catalog rows and verify availability before larger runs.
First test One small non-sensitive prompt Confirm endpoint, slug, latency, response format, and billed usage before production or repository-wide context.
Compatibility checklist

Confirm Cursor can actually use this setup

This table separates a real custom-endpoint setup from cases where the tool version, secret handling, or model field will make the connection fail.

Check Good state Avoid this
Required control Cursor exposes a custom base URL, custom provider, or compatible API host field. If Cursor hides endpoint controls, use the docs fallback instead of guessing hidden settings.
Endpoint shape OpenAI-compatible custom endpoint support is available and can be set to https://base.corvusllm.com/v1. Do not mix OpenAI-compatible setup values with another request format.
Secret handling The CorvusLLM key is stored in Cursor settings, credentials, or server environment variables. Do not paste API keys into public repositories, prompts, screenshots, or client-side production code.
Model slug The selected value is a public Claude slug from the model catalog. Do not use private upstream route names, provider-account names, or guessed model aliases.
First request A small non-sensitive prompt succeeds before larger project context, automation, or team rollout. Do not start with private repositories, regulated data, or high-volume automation before a small test.
Model source

Start from live Claude catalog rows

These examples are only starting points. The model catalog and data/models.json are the source of truth for availability, pricing, cache fields, and public slugs.

claude-haiku-4-5Check the live Claude catalog row claude-opus-4-5Check the live Claude catalog row claude-opus-4-6Check the live Claude catalog row All Claude rowsCompare the current family catalog Model docsSlug, pricing, and compatibility notes
Troubleshooting before support

Check the common Cursor failure points first

These checks turn common setup errors into crawlable answers and send private account cases to the right support path only after public checks are exhausted.

Symptom First check Public source
Custom base URL field is missing Confirm your Cursor version exposes custom provider controls. Cursor docs
Model not found Use a public Claude slug from the current catalog and avoid hidden aliases. Claude catalog
Authentication or balance error Confirm the delivered CorvusLLM key, account balance, and that the key was not pasted with extra spaces. Troubleshooting
Unexpected cost Estimate input, output, cache read, and cache write before sending large context or loops. Cost calculator
Timeout or provider unavailable Retry a tiny non-streaming prompt, then check customer-facing status before changing credentials. Service Status
Proof and guardrails

Verify price, status, and service boundary first

This is a commercial setup page, not a private account console. Use the supporting pages for exact pricing, operational status, and trust details.

Pricing trackerSource-linked public rates Cost calculatorEstimate token usage before top-up Trust CenterAffiliation, data, support, and limits Verify CorvusLLMSkeptical buyer checklist Service StatusCustomer-facing health checks Cursor docsExact setup guide
Data handling warning

Do not send sensitive or regulated data through shared API proxies. CorvusLLM forwards prompts to upstream model providers for processing and keeps request metadata for billing, abuse prevention, and support diagnostics.

Common questions

Before using Claude in Cursor

These answers keep setup pages useful for search, AI answers, and first-time buyers without replacing the exact docs.

Which base URL should Cursor use for Claude?

Cursor should use https://base.corvusllm.com/v1 for this Claude setup path. Check the dedicated docs page when the tool changes its custom-provider settings.

Do I need my own Anthropic account first?

No. This page is for using a CorvusLLM prepaid key as an independent access layer. CorvusLLM is not affiliated with Anthropic, and exact supported rows should be verified in the public catalog.

Can I paste sensitive Cursor project data into this setup?

Do not send sensitive or regulated data through shared API proxies without a separate risk review. Start with non-sensitive tests and move larger workloads only after checking trust, status, and billing behavior.

Related setup pages

Compare nearby tool and provider paths

Provider+tool pages are linked together so developers and crawlers can move from a broad query to the exact setup route.

Claude API for Claude CodeClaude setup for Claude Code Claude API for Open WebUIClaude setup for Open WebUI Claude API for OpenAI SDKsClaude setup for SDKs Claude API for n8nClaude setup for n8n Claude API for ChatBoxClaude setup for ChatBox GPT API for CursorGPT setup for Cursor GLM API for CursorGLM setup for Cursor GPT API for Open WebUIGPT setup for Open WebUI GPT API for OpenAI SDKsGPT setup for SDKs GPT API for n8nGPT setup for n8n GPT API for WindsurfGPT setup for Windsurf GPT API for ChatBoxGPT setup for ChatBox

Use a small Claude test in Cursor first.

Confirm base URL, key, public slug, latency, output quality, and billed usage before larger prompts, repository context, or automated workflows.

Get API Key All Tool Setups
Topic map

Continue with the right source

Compare the setup path, model catalog, pricing proof, and trust pages before you choose an endpoint.

landing AI API for Coding Agents CorvusLLM can fit coding-agent workflows when the user wants one prepaid key. landing AI API for Open WebUI Teams CorvusLLM can fit Open WebUI teams that need a custom OpenAI-compatible backend, a prepaid balance model. landing AI API for n8n Automation CorvusLLM can fit n8n automation when workflows need explicit HTTP request configuration, prepaid usage, public model slugs. landing AI API for App Prototyping CorvusLLM can fit app prototyping when the goal is to test an AI feature quickly with OpenAI-compatible SDKs. landing AI API for Cost-Sensitive Workloads CorvusLLM can fit cost-sensitive workloads when the user can estimate token volume, avoid sensitive data. landing AI API for Multi-Model Routing CorvusLLM can fit multi-model routing when the user wants one prepaid key for supported public catalog model families. landing Claude API Pricing Comparison CorvusLLM lists public Claude-family rows at 35% of tracked official input, output, cache-read. landing GPT API Pricing Comparison CorvusLLM lists public GPT-family rows through an OpenAI-compatible access layer with public prepaid rates derived from tracke. landing GLM API Pricing Comparison CorvusLLM lists public GLM-family rows for buyers who want cost-sensitive API options, but exact row availability. landing AI API Cache Token Pricing Cache-heavy requests can cost very differently from short prompts because cache read and cache write fields may dominate the b. landing OpenAI-Compatible AI API Proxy CorvusLLM provides an independent OpenAI-compatible AI API proxy for buyers who need prepaid balance, setup docs. landing AI API for Cursor CorvusLLM can be used in Cursor builds that expose custom provider fields; this page explains the commercial fit. landing Claude, GPT & GLM API CorvusLLM offers one independent endpoint for supported Claude, GPT, and GLM family access. landing Bulk AI API Access The bulk AI API page is for teams, agencies, and automation buyers who can describe expected usage, model families, key needs.