LATEST NEWS
SELECTED FOR YOU

WhatsApp begins testing third-party AI agents, up to five per account

ByRanda MosesRanda Moses 2 mins read
WhatsApp begins testing third-party AI agents, up to five per account.
  • WhatsApp is testing a feature letting users connect up to five outside AI agents via a generated API key.
  • Each agent sees only the chat it sits in and cannot join group conversations.
  • These agent chats are not end-to-end encrypted because they route through a Meta service.

WhatsApp is beginning to usher in a feature that allows outside AI agents to join the app for a dedicated chat, with no single agent linked to a specific AI provider. The code appeared in the Android beta on September 4.

Millions of people are already running agents on rival apps like Telegram. Now, AI labs will have the ability to tap into WhatsApp’s user base with a simple API key.

An API key links an external AI agent

The connection is based on a key that the user generates in WhatsApp. With the New Agents entry in settings, a user can set up an agent, name it, and choose a profile photo.

WhatsApp then yields an API key, which the user then pastes into the platform where the agent is hosted to complete the link.

Users can create a new key at any time to cut and reestablish the connection. WhatsApp allows up to five agents connected to an account.

WhatsApp begins testing third-party AI agents, up to five per account.
WABetaInfo’s screenshots of WhatsApp’s agent setup flow. Source: WABetaInfo.

Each agent runs in its own one-on-one chat, detached from the rest of the account. It only reads messages in its thread and has no access to any other chats, contacts, or media files.

For now, agents are confined to one-on-one chats and can’t be dropped into groups or communities. WhatsApp also warns that AI agent answers may be inaccurate, and users should verify anything it tells them.

Removing an agent disconnects it but does not erase the thread. Old messages stay visible, though reconnecting later means starting over with a new key.

Agent chats skip WhatsApp’s end-to-end encryption

Messages with external AI agents are not end-to-end encrypted since they go through a Meta service that acts on behalf of the agent’s developer.

When the service is running, the developer can see the messages. Meta says this does not expose the chats to unrelated third parties and that normal calls and messages between people remain fully encrypted.

The move flips WhatsApp’s earlier stance. With the rise of agent platforms, it was found that connecting a bot to a WhatsApp account could get that account banned.

Telegram embraced the trend and became a popular home to talk to cloud and local agents.

The rollout is currently restricted to a small pool of Android users in select countries on beta version 2.26.35.3, with a wider release expected over the coming months and no firm date for a global launch.

Meta is now directly vying with Google, Anthropic, and OpenAI, as the company launched a business agent in June that handles payments, bookings, and sales across WhatsApp, Instagram, and Messenger, as well as a platform for firms to build custom agents, reported Cryptopolitan.

The plan was to move beyond “rule-based automations” and have the software do the work itself, said Meta’s own product head, Naomi Gleit.

The smartest crypto minds already read our newsletter. Want in? Join them.

FAQs

How many AI agents can a WhatsApp user connect?

WhatsApp supports up to five third-party agents per account, each managed and chatted with in its own individual conversation.

Are chats with WhatsApp's third-party agents end-to-end encrypted?

No. Because the messages are handled by a Meta service on behalf of the agent's developer, they are not end-to-end encrypted, and the developer can access them.

Who can use the feature right now?

The feature is live for a limited number of Android users in select countries on beta version 2.26.35.3. A wider rollout is expected over the coming months, and there is no announced global launch date.

Share this article

Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.

Randa Moses

Randa Moses

Randa Moses is an editor and reporter at Cryptopolitan covering tech, AI, robotics, crypto, scams, and hacks. She has worked in the crypto space since 2017. She held roles at Forward Protocol, AmaZix, and Cryptosomniac. Randa holds a degree in Electrical and Electronics Engineering from the University of Bradford.

MORE … NEWS