CorvusLLM Docs - API Setup, Models and Billing

Start from zero

If you are new, start with a CorvusLLM key, a browser, and one tool you want to connect. Coding setups usually also need Git, Node.js, Python, or Docker depending on the guide you choose.

What to install	Official source	Quick check
CorvusLLM key	Pricing and checkout	Needed for every request.
Developer basics	Git install guide, Node.js download, Python downloads	Install only what your chosen guide requires.
Container apps	Docker Desktop docs	Needed for local Open WebUI only.

Popular setup guides

These crawlable links mirror the most common paths from the searchable docs UI, so users and search engines can reach the right setup guide without relying on JavaScript-rendered cards.

OpenAI SDKs →Python, Node, fetch, and curl examples for the OpenAI-compatible /v1 endpoint. Claude Code →Anthropic-native setup using /anthropic, the correct auth variable, and Claude slugs. Open WebUI →Where to paste the base URL, key, and starter model in a self-hosted chat UI. Cursor →Use CorvusLLM only when your Cursor build exposes a custom OpenAI-compatible endpoint field.

Which endpoint should I use?

/v1 for OpenAI-compatible tools

Use https://base.corvusllm.com/v1 for OpenAI SDKs, curl, fetch, Open WebUI, ChatBox, n8n, and most custom apps.

/anthropic for Claude-native workflows

Use https://base.corvusllm.com/anthropic for Claude Code or tools that expect Anthropic-native request handling.

Start with the right guide

Use the most direct path for what you are actually trying to do. This removes the duplicate "read everything first" feeling and gets you into the correct page faster.

The four rules that matter most

Most tools use /v1

SDKs, HTTP, Open WebUI, ChatBox, and most custom integrations start on https://base.corvusllm.com/v1.

Claude Code uses /anthropic

Do not point Claude Code at the OpenAI-compatible path. Use https://base.corvusllm.com/anthropic.

Copy the canonical slug

Use the short customer slug shown in the docs, not a guessed alias or internal long-form variant.

Balance lives on the same key

Top-ups credit the same customer key again. Out-of-balance stops usage instead of silently overspending.

Most-used environments

These are the common first stops. The full compatibility table still lives in the environment overview, but the home page only shows the setups people usually need first.

Model families

Use this section to orient quickly. The full canonical slug reference stays in the dedicated models page, where the exact table is easier to scan.

Reference pages for exact values

Quickstart → The shortest path from delivered key to a successful request and visible dashboard balance. Environment overview → Use this when you need to know which path and auth field your tool actually expects. Models & slugs → Copy the canonical slug instead of guessing aliases or long internal variants. Billing, Balance & Cache → Open the exact billing behavior for top-ups, zero balance, and cache pricing. API overview → Check base URL shape, compatible paths, authentication, and request basics. Before you buy → Understand proxy trade-offs, data boundaries, payment behavior, and when direct provider access is better. After you buy → What to verify after delivery before you connect bigger workflows. ChatBox setup → Configure a custom OpenAI-compatible provider with the CorvusLLM base URL and key. n8n setup → Use the HTTP Request node with explicit URL, bearer auth, and JSON request body. Windsurf setup → Use CorvusLLM only when your Windsurf build exposes a custom provider override. Troubleshooting errors → Diagnose auth, model-not-found, balance, and routing errors without guessing. Cost calculator → Estimate input, output, and cache-heavy usage before adding prepaid balance. Pricing tracker → Compare CorvusLLM rates with source-linked official provider pricing references.

Troubleshooting deep dives

Use these focused error pages when you already know the symptom and need the shortest safe diagnostic path.

Model not found →Fix unknown model, no such model, 404 model, and invalid slug errors. Invalid API key →Fix unauthorized, 401, 403, missing Bearer token, and secret-profile mistakes. Wrong base URL →Fix double /v1, wrong endpoint, and Claude Code /anthropic routing mistakes. Out of balance →Fix 429, insufficient credits, top-up, and cache-heavy billing issues. Provider unavailable →Fix temporary upstream route failures and model-family outage symptoms. Request timeout →Fix long request, gateway timeout, streaming, and large-context latency failures. Image input errors →Fix vision payload, base64 image, attachment, and multimodal request errors. Tool creation failed →Fix tool calls, file writes, workspace patches, schema issues, and agent permissions. Cache cost spikes →Fix hidden context, cache read/write, continue prompts, and unexpected balance drops.

Compare before top-up

Use these pages when you are choosing a model family, estimating cost, or checking whether a third-party proxy is the right fit for your workload.

Model catalog →Start from the public model index before copying a slug into a client. OpenAI-compatible API →Understand the buyer-facing API proxy page and service boundaries. Cursor API setup →Use the dedicated buyer page for Cursor custom base URL searches. Claude, GPT, Gemini & DeepSeek API →Compare multi-family API access through one prepaid key. Bulk API access →Read the team and agency route before larger prepaid purchases. Trust Center →Read data handling, support limits, refund notes, and affiliation boundaries. Get API key →Open the public prepaid key section after setup and risk checks are clear.

Set up CorvusLLM without guessing.