Safety & Privacy
How Convos assistants handle your messages and protect your privacy.
Convos the app is open source, end-to-end encrypted, and collects no data — not even contact info. There are no accounts, no phone numbers, no usernames. You're identified by a key stored on your device, and every conversation gives you a fresh identity. Nobody — including us — can read your messages or connect your conversations to each other.
Convos assistants — the ones you add through the Convos app or this website — are a separate, hosted service layered on top. When you add one to a group conversation, that assistant processes your messages in order to respond and take actions. This data is processed on infrastructure operated by XMTP Labs.
You can also add any AI agent to a Convos chat — third-party agents have their own privacy practices. Everything on this page applies to Convos assistants specifically.
Sensitive information
Assistants can help with sensitive tasks — that's part of the point. But remember that anything you share is processed by the third-party services listed above, and visible to everyone in the group. Be intentional about sharing things like passwords, financial account numbers, government IDs, or medical records.
What the assistant can see
The assistant reads messages in your group conversation in order to respond and take actions.
Messages in the group. The assistant can see messages sent by anyone in the conversation. It uses these to understand context and respond helpfully.
Only that group. Each assistant is fully isolated. It cannot see other conversations, contacts, profiles, or any data outside the group it was added to. Because Convos gives you a fresh identity in every conversation, the assistant has no way to correlate you across groups — even if you use assistants in multiple conversations.
Nothing else. No cross-pollination between groups. No access to your other chats. No shared memory across different conversations. No persistent profile of you.
One conversation per assistant
Every assistant is bound to a single conversation. It is created fresh with its own credentials, memory, and context — none of which are shared with any other assistant.
When the conversation ends, the assistant and all its data are destroyed. No residual data, no lingering access, no archive.
What each assistant is made of
Every assistant runs on OpenClaw, an open-source agent runtime. When you add an assistant to a conversation, it gets provisioned with its own isolated set of credentials for the following services:
| Service | Provider | What it does |
|---|---|---|
| LLM routing | OpenRouter | Routes prompts to the model provider best suited for each task. |
| AI models | Anthropic, OpenAI, Google, Meta, DeepSeek, and others | Your messages are sent to whichever model the assistant selects — Claude, GPT, Gemini, Llama, etc. Each provider has its own data policies. |
| Web search | Perplexity via OpenRouter | Real-time web lookups. Queries are sent to Perplexity's Sonar model through OpenRouter. |
| AgentMail | Each assistant gets a unique inbox. Emails sent and received by the assistant pass through AgentMail's API. | |
| SMS & phone | Telnyx | Each assistant can be assigned a US phone number. Text messages and calls are routed through Telnyx. |
| Crypto wallet | Bankr | Onchain wallet for trading and transfers. Transaction requests are processed through Bankr's API. |
| Web browsing | Chromium (local) | A headless browser running on the same server as the assistant. No third-party service involved. |
| Messaging | XMTP | The assistant communicates with your group over the XMTP network, the same protocol Convos uses for all messages. |
Each assistant gets its own API keys for these services, created at startup and destroyed when the conversation ends. No keys are shared across assistants. The assistant runtime, hosting infrastructure, and all provisioning are operated by XMTP Labs on Railway.
You are in control
Remove the assistant anytime. Any group member can remove the assistant from the conversation at any time.
Explode the convo. Delete the conversation entirely and the assistant plus all its data are permanently destroyed.
Keep it humans only. Any conversation can be just people — fully end-to-end encrypted, no AI involved. When you want an assistant, add one. When you don't, don't. Both coexist naturally, and the choice is always yours.
Bottom line: Convos = private, encrypted, open source, no data collected. Convos assistants = a hosted service that processes your messages through third-party AI models and tools. Each assistant is isolated to one conversation, gets its own credentials, and is destroyed — along with all its data — when the conversation ends.