Tools & keys · AI

Cloudflare Workers AI setup

Open models on Cloudflare's network, with a daily free allowance.

Optional
Updated
Free tier
Cloudflare's free plan includes a daily allowance of AI compute (10,000 "neurons" a day at the time of writing), with no card required.
checked September 2026
Larvabot's cap
Nothing to cap: rate limits just hand the job to the next provider.
Fallback position
#4 of 7 for AI. See the order

What it does

Workers AI runs open models such as Meta's Llama on Cloudflare's GPUs, with an OpenAI-compatible API.

  • A reliable AI fallback, running Llama 3.3 70B by default

Get set up

  1. Copy your Account ID from the Cloudflare dashboard
  2. Create an API token from the Workers AI template
  3. Paste both in Settings → AI
Open Workers AI

Where it goes

In the dashboard, open Settings and find the AI section. Secret values are encrypted before they're saved and only their last four characters are ever shown again.

Settings → AI
Cloudflare account ID
CLOUDFLARE_ACCOUNT_ID
…
Cloudflare Workers AI tokenencrypted
CLOUDFLARE_AI_TOKEN
••••••••••••
Cloudflare model
CLOUDFLARE_AI_MODEL
@cf/meta/llama-3.3-70b-instruct-fp8-fast

Staying free

Larvabot moves on to the next provider when the daily allowance is used up, and asks for longer answers than Cloudflare's short default so drafts aren't cut off.

How the quota guard works →

Fallback order

For AI, Larvabot tries these in order and skips any that aren't set up, are out of free quota or are busy. AI providers also have to pass the quality check on the Jobs & providers page.

  1. 1Google Gemini
  2. 2Groq
  3. 3OpenRouter (:free models)
  4. 4Cloudflare Workers AI
  5. 5Cerebras
  6. 6Mistral
  7. 7Custom OpenAI-compatible (e.g. Ollama)

More AI tools

Curious why Cloudflare Workers AI made the cut? Cloudflare Workers AI in the free stack →

Keep going