How to create a Llama API Key for Termerly

A simple guide to connect your Termerly account to Llama and generate policy drafts using Meta's models

Why does Termerly need your Llama API Key?

Termerly generates policy drafts (privacy, terms, cookies, etc.) using artificial intelligence models. For those drafts to be written with Llama (Meta's models), Termerly needs an API Key that lets it connect to whichever service hosts those models.

An API Key is like a personal key that lets Termerly talk to Llama on your behalf. You generate it once, paste it into Termerly's settings, and from that moment on you can generate your drafts with Llama.

Important note about Llama. Unlike OpenAI or Anthropic, Llama doesn't have a single provider. Meta publishes the models as open source, so there are several ways to obtain a compatible API Key. The simplest is the official Meta Llama API, but there are alternatives if that one doesn't fit you, which we'll cover at the end of this guide.

Recommended path: the official Meta Llama API

The official Llama API is the "natural" path if you want to use the models directly from Meta, with its own dashboard, latest models, and preview-period limits. As of writing, the API is in limited preview: access is free and granted through a waitlist, mainly aimed at developers with verified US accounts.

If you're in a hurry and don't want to wait for the waitlist, jump to the "Alternatives" section at the end of this guide: third-party providers serve Llama commercially with no delays.

What you'll need on hand

A Meta developer account (or a personal Facebook or Instagram account to sign up with).
A valid email address.
A modern browser (Chrome, Firefox, Edge, or Safari).
Access from a supported region (currently mainly the US).

Estimated time: 5 minutes to register and fill the form; actual access can take from a few minutes to several weeks.

Good news: during the preview period you don't need a credit card. Access is free while this phase lasts. When the API enters general availability, the usual billing system will kick in.

Step 1: Create your account and join the waitlist

Open your browser and go to https://llama.developer.meta.com.

Click "Join the waitlist" or "Sign up". You can register with an email or connect your existing Meta account.

Suggested image: the llama.developer.meta.com homepage with the "Join waitlist" button visible.

After registering, Meta will send you a confirmation email. Then you'll have to fill in a brief form explaining what you want to use the API for:

The use case (something like "generating policy drafts for my website" works).
The type of application.
The approximate usage volume.

This information is used to prioritize approvals during the preview.

The approval timeframe varies: in some cases it's nearly instant, in others it can take several days or weeks. You'll get an email when access is granted.

Step 2: Access the dashboard once approved

When your account is approved, you'll receive an email with the access link. From there, you can enter the dashboard at https://llama.developer.meta.com.

The dashboard offers:

An interactive playground to try the models directly from your browser.
The API Keys section.
Documentation and examples.
Usage and limits info.

Suggested image: the main dashboard view of the Llama API after signing in.

Step 3: Generate your API Key

Inside the dashboard, find the "API Keys" section in the side menu.

Click the "Create API key" button (or the "+" icon). A small window will pop up where you can:

Assign a name: type "Termerly". That way, if you create more keys later for other tools, you'll know which one is for Termerly.
Confirm the creation.

Suggested image: the "Create API key" dialog with the Name field filled in as "Termerly".

Click "Create". You'll see a long key with a format like this: LLM|XXXXX|XXXXXXXX.... That's your API Key.

Copy it RIGHT NOW! Meta only shows it to you once. If you close the window without copying it, you won't be able to recover it and will have to create another. Click the "Copy" button next to the key.

Step 4: Paste it into Termerly

Go back to Termerly and into your dashboard:

Sign in to your Termerly account.
In the side menu, go to Settings → Account.
Scroll down to the Llama (Meta) section.
Paste the key you just copied into the "Paste your API key" field.
Click the "Save" button.

Suggested image: the Llama (Meta) section in Termerly, inside Settings → Account, with the "Paste your API key" field filled in.

If the key is valid, you'll see the field's status change from "Not connected" (red) to "Connected" (blue). From that moment on, Termerly can generate drafts with Llama models on your behalf.

Your key is safe. Termerly stores it encrypted with AES-256-GCM and never shows it back to you in plain text. If you ever need to see it again, you'll have to go back to Meta and generate a new one.

Step 5: Verify the connection works

Termerly has a button labeled "Test connection" right below the API Key field. Clicking it is the fastest way to know if everything is right:

If the connection works, you'll see a green check and a "Last used" date will appear under the field.
If the connection fails, Termerly will tell you what happened (invalid key, region not supported, etc.).

Suggested image: the "Test connection" button in Termerly with a successful result and the "Last used" timestamp visible.

From here you can start generating policy drafts in Termerly with Llama models.

Step 6: Take care of your API Key

Your API Key is a personal credential: even though usage is free during the preview period, a leaked key can drain your daily limits and block service for your own application. When the API enters general availability it can also generate charges. So, from day one:

Don't paste it in public chats or share it via email, social media, forums, or videos.
Don't post it where others can see it (screenshots on social media, for example).
Create a different key for each tool. The one you generated for Termerly should only be used in Termerly.
If you suspect it has leaked, run to the Llama dashboard and click "Revoke" on that key. Then generate a new one and paste it back in Termerly using the "Update" button.
Check the "Usage" section in the dashboard every now and then to spot unusual consumption.

If you think of your API Key as your bank card PIN, you'll be on the right track. Same care, same discretion.

Things that can go wrong and how to fix them

What you see in Termerly	What it means	How to fix it
"Not connected" after clicking Save	The key is wrongly pasted or invalid	Generate the key again on the Llama dashboard and paste it back, no leading or trailing spaces.
"403 Forbidden" or "access denied" when testing	Your account doesn't have preview access yet	Check your waitlist status on the Meta dashboard.
"Region not supported"	You're calling from a region not enabled	The preview is limited to the US. Consider using a third-party provider (see alternatives below).
"Rate limit exceeded" when generating a draft	Too many requests in a row for the preview limits	Wait a minute and try again.
"Invalid API key" or "401 Unauthorized"	You pasted the key wrong or already revoked it	Generate a new one on the Llama dashboard and click "Update" in Termerly.

Alternatives if you don't want to wait for the waitlist

Since Llama is an open-weights model, you're not bound to the official Meta API. Several third-party providers serve the same models commercially, with no delays and from any region. The most popular ones are:

Groq (groq.com): ultra-fast inference.
Cerebras (cerebras.ai): another low-latency option.
Fireworks AI (fireworks.ai): broad catalog with good pricing.
Together AI (together.ai): varied open source models.
OpenRouter (openrouter.ai): a single API routing to multiple providers.
AWS Bedrock and Azure AI Foundry: ideal if you're already inside those cloud ecosystems.

The process to get a key is very similar across all of them: you sign up, add a card (typical minimum charges are very low), generate an API Key, and paste it into Termerly just like you would with the official one. Each will give you a key in its own format, but the flow in Termerly is identical.

Recommendation: if you've never worked with these providers, start with Groq. It's free to get started, very fast, and tends to approve accounts instantly.

Wrapping up

Connecting Termerly to Llama isn't as complicated as it sounds, but it has a twist compared to OpenAI or Anthropic: you have to decide which provider you want to get the key from. The "natural" option is the official Meta Llama API, free during the preview but with a waitlist and regional restrictions. If you're in a hurry, third-party providers like Groq, Cerebras, or Fireworks AI are a perfectly valid alternative and tend to have better latency.

Whichever path you choose, the rest of the process in Termerly is the same: paste the key in Settings → Account → Llama (Meta), click Save, and start generating drafts. If you ever get lost, come back to this guide and re-read the relevant step.

Stay up to date

Subscribe to get notified when new articles are published.