Live on EvoLinkMODEL · claude-fable-5

Claude Fable 5 APIis live

Call Claude Fable 5 through the EvoLink unified API gateway. Use one API key to bring Fable 5 into agents, code review, complex planning, and long-context tasks.

Get API key View example

One API key · OpenAI-compatible · Model details and pricing

Status

● HTTP 200 verified

Model ID

claude-fable-5

Endpoint

/v1/chat/completions

SDK

OpenAI-compatible

01 / Quickstart

Run Claude Fable 5 API
in one minute

Start with one human action, then choose the integration path that fits your workflow.

Step 1

Create or copy your EvoLink API key

Open the API key page once, save the key as EVOLINK_API_KEY, and come back here. If you already have a key, go straight to the next step.

Get API key

Then choose how to integrate

Path A / Agent setup

Use an agent to integrate it

Copy this prompt into Cursor, Claude Code, Codex, or another coding agent, add your API key at the end, and let it wire Fable 5 into your project.

Agent prompt

Help me integrate Claude Fable 5 API into the current project.

Key facts:
- Model: claude-fable-5
- Base URL: https://direct.evolink.ai/v1 (OpenAI-compatible chat completions, Bearer auth)
- Model details: https://evolink.ai/claude-fable-5
- Examples: https://github.com/Evolink-AI/awesome-claude-fable-5
- Only if this project already uses the Anthropic SDK, keep it and use base URL https://direct.evolink.ai (Anthropic Messages protocol) instead — reference: https://docs.evolink.ai/en/api-manual/language-series/claude/claude-messages-api.md

What to do:
- Inspect how this project currently calls LLMs
- If the project has no LLM calls yet, create a minimal standalone example script instead of restructuring anything
- Make the smallest reasonable code change, using model claude-fable-5
- Add or describe a local test method
- Summarize what changed and any setup steps I need to run

My EvoLink API key (use it as the Bearer token): <paste it here>

Prefer not to paste your key into the agent? Use an environment variable

With this option, delete the key line at the end of the prompt. The command below is generated locally in your browser and the agent reads EVOLINK_API_KEY from your environment, so the key never enters the chat.

Run this before starting the agent

export EVOLINK_API_KEY='<your-evolink-api-key>'

The command only applies to the current terminal window — run it again in new terminals.

Path B / Manual setup

Change the code yourself

Prefer to inspect the request directly? The examples below show common starting points, not every supported language.

# one key, OpenAI-compatible. Change the model only.
curl https://direct.evolink.ai/v1/chat/completions \
  -H "Authorization: Bearer $EVOLINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-fable-5",
    "messages": [
      { "role": "user", "content": "Review this implementation plan and point out the risks most likely to cause rework." }
    ],
    "max_tokens": 1024
  }'

These are the three shortest examples. Other languages can use the same endpoint, model, and OpenAI-compatible request shape.

Open full docs ↗

02 / Real tasks

Test with real work

Do not judge Fable 5 with simple Q&A only. Put it into tasks that are complex, judgment-heavy, and costly to redo.

Agent

Agent task planning

Ask the model to break down the goal, identify risks, and plan the steps before an agent executes and verifies the work.

Good test case →Review

Large PR review

Use it to review implementation plans, potential regressions, edge cases, and high-risk changes before merge.

Good test case →Infra

Incident diagnosis

Give it logs, context, and previous debugging steps so it can suggest likely causes and next validation paths.

Good test case →Codebase

Repository audit

Let it read project structure, critical modules, and constraints, then produce risks, refactor ideas, and migration plans.

Good test case →Research

Long-context analysis

Use it for long documents, long conversations, complex requirements, and cross-file context.

Good test case →

See more Fable 5 examples

03 / Why EvoLink

Choose EvoLink for
Claude Fable 5 API

If your product already uses an OpenAI SDK or compatible API, point requests to EvoLink and set model to claude-fable-5. API keys, usage, and billing stay in one place.

01
One API key
Create one EvoLink API key and start testing without managing multiple provider credentials.
02
OpenAI-compatible calls
Keep your existing request shape and set model to claude-fable-5.
03
Switch between models
Use Fable 5 for more complex work while keeping other models for default traffic.
04
Centralized usage
Track requests, spend, and model distribution from one EvoLink console.

04 / Learn more

Explore model details

Check pricing and parameters on the model page, compare Claude models before choosing one, or open the case library for real workflow examples.

Model details

Pricing, parameters, and API details

Review Fable 5 pricing, parameters, API availability, and integration notes.

View model details ↗

Claude family

Compare Claude models

Compare Fable 5 with Opus, Sonnet, Haiku, and other Claude models before choosing your production route.

Compare Claude models ↗

GitHub examples

Examples and prompts

Browse practical Fable 5 examples for agents, coding, long-context work, and automation.

View GitHub cases ↗

Ready when you are

Start calling Claude Fable 5 API

Get an API key, run your first request, and test Fable 5 in a real agent or coding workflow.

Get API key View docs