PRICED FOR PEOPLE
WHO ship, NOT
WHO PROCURE.

Yes, that is the core workflow. Enter your URL on the homepage to preview your bot first (no signup, no card). When you're ready, sign up, pick a plan, and saavos crawls the page, chunks the text, and builds a retrieval index in under a minute. You can add multiple URLs, a sitemap, PDFs, or pasted text. When your site changes, re-ingest the URL to update the bot.

No card to preview. Paste your URL on the homepage and you can talk to a bot trained on your site immediately — no signup, no card. You only enter payment details when you decide to install the widget on your real site.

Your bot runs on frontier Claude models from Anthropic. We pick the right model for each tier and upgrade it automatically as Anthropic ships better ones, so you never manage model versions.

Yes. Your knowledge sources are stored with row-level security — only your account can read or modify them. Chat logs are isolated per bot owner. We never use your data to train models.

If you sign up and decide within 7 days that saavos is not right for you, ping us in the dashboard chat and we will refund the charge in full — no questions asked.

Yes. Upgrade from your dashboard for an instant prorated charge. Downgrade or cancel at the end of your current billing period.

Yes. The Solo plan is $9/month and runs frontier Claude on every response. Solo unlocks 1,000 messages/month and one website — no per-message fees, no resolution charges. Builder ($24) runs the same model with higher limits, and there's a forever-free tier if you want to start without paying.

Copy the snippet from your dashboard — one <script> tag — and paste it before the closing </body> tag on your site. That is the entire integration. The widget loads asynchronously and works on any site that accepts custom HTML: WordPress, Webflow, Shopify, Framer, or raw HTML. No npm package, no API keys in your frontend, no configuration file.

Yes. Cancellation takes effect at the end of your current billing period. Your bot keeps working until then, and you never owe a penalty.

Typical first-token latency is under two seconds. Responses stream word-by-word so visitors see progress immediately.

Yes — conversation export (JSON or Markdown) is available on all plans from the dashboard.

SSO and custom hosting are on the roadmap. If you need them today, ping us in the dashboard chat — we can discuss a custom arrangement.

Your bot returns a friendly "limit reached" message and resumes at the start of your next billing period. No plan has per-message overage — no surprise bills, ever. If you keep hitting the cap, that is the signal to move up a tier.