Privacy
Plain-language policy · Last updated September 1, 2026
chanio exists because what you type to an AI is some of the most sensitive data you produce. This policy is written the way we'd want to read one: short, specific, and honest about what happens where.
The product is local-first
The chanio capture and compile pipeline writes to your own disk. Your channels, your knowledge graph, your morning briefs — they are files you own, on hardware you control. They are not uploaded to chanio servers, not used to train models, and not visible to us. Portability is the point: your graph travels with you, not with us.
What this website collects
If you join the beta list, we store the email address you give us, and we use it to contact you about chanio — nothing else. We don't sell it, share it, or enrich it. The conversation on /join is answered by a model through chanio's server. Once it names a channel, that channel and the transcript are saved — with your email if you leave one — so your beta invite can be personalized. Ask and we delete it. The Lab is bring-your-own-key: your API key stays in your browser and your requests go directly to the model provider you choose.
How we count visits
This site counts its own page views, first-party: the page, the referring site, the day, and whether the device is mobile or desktop. No cookie is set, no IP address is stored, no fingerprint is taken, and no third-party analytics script runs. Browsers that send Do Not Track are not counted. Raw rows are kept for 90 days; only the aggregates are readable, and they are public at /stats. Everything a visitor writes to this site — a page view, a beta-list email, the channels named on /join — goes through one rate-limited route on our server, never straight to a database.
What passes through our servers
Almost everything chanio does happens on your machine. These four things don't, and this is exactly what happens to each:
- The /join conversation. Answered by a model through chanio's server; the named channel and the transcript are saved as described above.
- /import's backfill and article lanes. Your API key and the conversations or articles you send ride through a chanio edge function to Anthropic. They are not logged and not stored.
- The /lab “Add URL” fetch. URLs you add are fetched by chanio's server and the page text is returned to you. Everything else in the Lab goes straight from your browser to Anthropic.
- The in-browser model on /try. A model file is downloaded into your tab. Nothing you type there leaves it.
Third parties
The services that touch any of the above, and what each one touches:
- Supabase — email and beta-list storage.
- Stripe — payments, when checkout opens. We never see your card number.
- Anthropic — the model behind /join and /import.
- harnesshealth.ai — the shared footer script every page loads; it is the same footer our sibling sites use.
- The browser-model CDN — where /try downloads its in-browser model file from.
What we never do
We never sell your data. We never train on your captured context. We never read your vault. If a future feature would ever require any of that to change, it will be opt-in, explained in plain language, and this page will change first.
Questions or deletion requests
Write to orthoblaino@gmail.com — or reply to any chanio email — and a human will answer. Beta-list removal and transcript deletion requests are honored promptly.