1. Local Execution
PicoClaw runs as a single Go process on your machine. All data processing happens locally. Your conversations, files, and settings never leave your device.
Last updated: February 10, 2026
• 100% Local Processing: All your data stays on your device. Period.
• Zero Tracking: No analytics, no telemetry, no spying.
• Open Source: MIT licensed - you can verify everything.
• You're in Control: Your data, your rules, your infrastructure.
PicoClaw is built on a fundamental principle: your data belongs to you. Unlike cloud-based AI services, PicoClaw runs entirely on your own hardware, ensuring complete data sovereignty and privacy.
We don't operate servers that collect your data. We don't have analytics tracking your usage. We can't see your conversations, your files, or your information - because we never receive them.
We don't collect names, emails, phone numbers, or any identifying information.
Your chats stay in your local SQLite database. We never see them.
No telemetry, no metrics, no tracking pixels. Your usage patterns are private.
Files you share with PicoClaw remain in your containers. We have zero access.
We don't fingerprint your device or collect hardware specifications.
No server logs, no connection tracking, no location data.
PicoClaw runs as a single Go process on your machine. All data processing happens locally. Your conversations, files, and settings never leave your device.
All data is stored locally in isolated directories. Only explicitly configured paths are accessible, providing an additional security layer.
Messages and state are stored in a local SQLite database on your filesystem. This database is yours - you can back it up, delete it, or migrate it as you see fit.
Each platform (Telegram, Discord, QQ, etc.) maintains isolated context. All data is stored locally and never synchronized to external servers.
While PicoClaw itself doesn't collect data, it integrates with external services that have their own privacy policies:
When you use PicoClaw, your prompts are sent to your configured LLM provider's API to generate responses. Each provider processes these requests according to their own privacy policy. Your API keys are stored locally in your environment variables.
Important: Most reputable LLM providers state they don't use API data to train models.
If you use messaging platform integrations (Telegram, Discord, QQ, DingTalk, etc.), messages are transmitted through their respective infrastructures and subject to their privacy policies. PicoClaw connects to these platforms using your credentials - no data is stored on our servers.
When PicoClaw performs web searches or fetches content on your behalf, those requests go directly from your device to the target websites. We don't proxy or log these requests.
Since PicoClaw runs on your infrastructure, you control the security:
Because PicoClaw is self-hosted and open source, you have complete control:
All data is in your local SQLite database and files
Simply delete the database file or entire installation
SQLite databases are portable and readable
MIT license allows complete customization
This documentation website is a static site hosted on a CDN. We don't use:
Your visit to this website generates standard web server logs (IP addresses, timestamps, pages visited) that are automatically deleted after 30 days. These logs are used solely for security and debugging purposes.
PicoClaw is designed for general audiences. Since we don't collect any personal information, there's no special consideration needed for children's data. However, parents should supervise their children's use of AI assistants and be aware of LLM providers' policies regarding age restrictions.
As PicoClaw is open source and self-hosted, you control when and if you update. If we make changes to this privacy policy, we'll update the "Last updated" date at the top of this page and note changes in the GitHub repository's changelog.