Bring your Anthropic key and power your WhatsApp agent with Claude models
Anthropic builds the Claude family of large language models — Claude Opus, Sonnet, and Haiku — known for strong reasoning, long context, and reliable tool use. The LunoGen Anthropic integration lets you bring your own Anthropic API key and power your agent with Claude models. You can pick different models for different tasks, enable prompt caching to cut costs, use the Messages API for structured tool calls, and run vision tasks on images sent in WhatsApp.
Many teams prefer Claude for long-form reasoning, agentic tool use, and safer default behavior. Anthropic also has a strong prompt caching feature that can cut repeated-context costs by 90%. Bringing your own Anthropic key gives you billing control, access to the latest Claude models, and the option to route heavy tasks to Opus and fast replies to Haiku — while LunoGen handles the plumbing.
In the Anthropic console, go to Settings → API Keys → Create Key. Name it 'LunoGen Agent' and copy the key.
In LunoGen, open Credentials → New → Anthropic and paste your API key. Optionally add your workspace ID.
Choose a reasoning model (Claude Opus recommended) and a fast model (Claude Haiku recommended). Sonnet is a good middle ground.
Turn on prompt caching. LunoGen automatically caches long system prompts and recent memory so repeated calls pay only for the delta.
Configure daily and monthly spend caps. LunoGen will pause new Claude calls and ping you on WhatsApp if a cap is reached.
Connect your Anthropic API key and LunoGen uses your account for all Claude calls. You control billing, model access, and rate limits — and gain access to the latest Claude releases as soon as they ship.
Switch my agent to use my Anthropic key and make Claude Opus the default
Validating Anthropic key and checking model access...
Route your hardest tasks — deep research, multi-step tool use, long planning — to Claude Opus. The agent detects complexity and automatically uses Opus when the prompt needs it.
I need you to plan the Q2 launch campaign end to end. Pull market data, draft the email series, pick a launch date, and write the press release.
This is a multi-step planning task. Routing to Claude Opus with tool use...
Quick user messages — greetings, simple questions, intent checks — get routed to Claude Haiku for near-instant, low-cost replies. Opus handles the thinking, Haiku handles the talking.
hey what's my meeting with priya tomorrow?
Quick query — routing to Claude Haiku...
LunoGen caches your long system prompt, memory context, and knowledge base on Anthropic's side. Repeated calls hit the cache and cost a fraction of normal tokens — often cutting bills by 70% or more.
Show me how much prompt caching is saving us this week
Pulling cache hit stats from Anthropic usage logs...
LunoGen uses Anthropic's Messages API with tool use for structured agent actions. Claude decides which tool to call, and the agent executes HTTP requests, sends WhatsApp messages, or writes files in one loop.
Check the weather in Tokyo and if it's raining, reschedule my 3 PM walk to tomorrow
Calling Messages API with tool use (weather + calendar tools)...
Send an image to the agent on WhatsApp and it runs Claude vision — reading receipts, describing screenshots, extracting handwritten notes, or checking a product photo for defects.
[sends a photo of a dinner receipt] can you log this to my expenses?
Running Claude vision on the receipt image...
Bring your own OpenAI key and power your LunoGen agent with GPT models
Ship code faster — PR reviews, triage, and release notes run by an AI agent
Run your Notion workspace from WhatsApp — create pages, query databases, and answer questions from your wiki
Turn any WhatsApp message into studio-quality voice with ElevenLabs AI
Spin up an AI agent in under 2 minutes, plug in your Anthropic credentials, and start running your workflows from WhatsApp.