Back

AI Policy

Effective date: April 27, 2026

This policy explains how Canopy uses artificial intelligence, what data is sent to the AI, what its outputs are and are not, and the controls you have over them. It supplements — and where there is conflict, defers to — our Terms of Service and Privacy Policy.

1. What "Canopy AI" is

Canopy AI is a chat-style assistant built on top of Anthropic's Claude large language model. It is designed to help you understand market context, your portfolio, and the news that moves it. It is a conversational interface to information — nothing more.

2. The model we use

All AI responses in Canopy are generated by Anthropic Claude, accessed through Anthropic's commercial API. Claude is a third-party service operated by Anthropic, PBC. Your use of Canopy AI is therefore also subject to Anthropic's Acceptable Use Policy.

We may switch underlying models or providers in the future to improve quality, latency, or cost. When that happens, we will update this policy and continue to disclose the active provider.

3. What we send to the AI

When you submit a message to Canopy AI, the following is sent to Anthropic's servers as part of the API request:

• The text of your message
• Up to the last 5 messages of your current conversation, for context
• Tickers in your portfolio and watchlist, plus limited position context such as share count, average cost, current price, and unrealized gain/loss when needed to answer your request
• An anonymized user ID, used by us for rate-limiting and abuse prevention only

4. What we never send

We deliberately do not send the following to the AI:

• Your email address, password, or authentication tokens
• Your raw transaction history, notes, full account balance, billing details, or authentication tokens
• Your full real name, billing address, or payment details
• Any data from your device beyond what you type and the limited portfolio context above

5. How long Anthropic keeps your data

Per Anthropic's commercial terms, API inputs and outputs are not used to train Anthropic's models. Anthropic retains API request data for a short period (currently up to 30 days) for trust, safety, and abuse prevention, after which it is automatically deleted. See Anthropic's Privacy Policy for the authoritative version.

6. How long Canopy keeps your AI history

By default, Canopy AI conversations are ephemeral on your device — they live in memory only and are lost when you close the app or reload. If you opt in via Settings → Canopy AI → Save conversations, your last 40 messages are stored locally on your device. We also keep server-side AI request and response logs for reliability, abuse prevention, and safety review; these logs are deleted with your account unless retention is legally required.

7. AI outputs are imperfect

Large language models can:

• Hallucinate. Generate confident, well-formed text that is factually wrong, including fabricated quotes, statistics, or earnings dates.
• Be out of date. The model has a knowledge cutoff and may not know about events that happened after that date. We try to compensate with live market data, but coverage is incomplete.
• Misread context. Especially with thinly traded tickers, ambiguous symbols, or non-US exchanges.
• Reflect bias from training data. No model is neutral, and ours is not an exception.

Treat every AI response the way you would treat a confident stranger on the internet — interesting, occasionally useful, never the final word. Always verify before acting.

8. Content moderation and reporting

Canopy AI is filtered both by Anthropic's built-in safety systems and by Canopy-side guardrails that prevent specific buy/sell recommendations, price predictions, and instructions for unlawful activity (such as market manipulation, insider trading, or evading disclosure obligations).

If you encounter a response you believe is harmful, defamatory, illegal, or otherwise objectionable, please report it by emailing support@faldar.ca with a copy of the prompt and response. We review reports within 7 business days and remove offending behavior at the prompt or model layer.

9. No personalized investment advice

Canopy AI is configured to refuse certain requests and to qualify others. In particular, it will:

• Refuse to tell you to buy, sell, or hold a specific security
• Refuse to predict a future price or return
• Refuse to construct or recommend a personalized portfolio allocation as if it were advice
• Decline to opine on tax strategy beyond stating that you should consult a licensed professional

If you push it to do these things anyway, treat the response as a creative-writing exercise, not financial advice. You alone are responsible for any trade you place.

10. Age requirement

You must meet the age requirement shown in the App Store listing for your region to use Canopy AI features. Canopy is not designed for children and we do not knowingly collect data from anyone under 13.

11. Your controls

• Skip AI entirely. Canopy works as a portfolio tracker without ever invoking the orb. AI features are gated behind a Pro subscription on the Free tier and never auto-engage.
• Save toggle. Turn local conversation persistence on or off in Settings → Canopy AI.
• Clear cache. Settings → Danger Zone → Clear Cache wipes all local AI history.
• Delete account. Settings → Account → Delete Account removes your Canopy account and any associated server-side state.

12. Trademarks

Claude® and Anthropic® are trademarks of Anthropic, PBC. Canopy is not affiliated with, endorsed by, or sponsored by Anthropic. Use of Anthropic's API is subject to Anthropic's commercial terms.

13. Changes to this policy

We will update this page when we change AI providers, change what data is sent, or change retention. Material changes will be surfaced in the app before they take effect. Continued use of AI features after a change constitutes acceptance.

14. Contact

For questions about this AI Policy or to report a problem with an AI response: support@faldar.ca