Karte
A public profile that answers back
About Karte
Karte is an open-source AI link-in-bio that turns a public profile into something visitors can talk to. A creator or independent operator publishes links, projects, timeline entries and profile context on a karte.cc page, and visitors can either browse it like a normal profile or ask a question that the built-in assistant answers from the content the owner chose to publish. Karte calls this a public inbound assistant: instead of guessing which button holds the answer, a potential client, collaborator or another AI agent can ask directly and then get routed to the right link or contact path.
The idea is that people should understand who you are before they ask for your time. Answers are grounded in the owner's public content and supplied memory, and when the source does not support an answer the assistant falls back to a bounded public-bio response rather than inventing private knowledge. Visitors do not need an account, while owners sign in with Google to claim and manage the page. Karte is free to use today with no billing in the current product, and the operator states plainly that the long-term commercial model is not committed. The full source code is published on GitHub under the MIT license, and the project is maintained by the Significant Hobbies studio, with a public changelog documenting recent work on chat reliability and inbound handling.
You start by drafting a page on karte.cc before signing in: choose a username, write a bio, add links and pick a theme. When you are ready to save and claim the page, you sign in with Google and move into the owner dashboard.
In the dashboard you manage profile sections, links, projects, timeline entries and appearance, and you add context that the chat assistant can use. You configure which visitor interactions are on - chat, contact forms or an email gate - and review exactly what will be public before you publish.
Once the page is live, you share the profile URL. Visitors can scroll your work or ask questions such as which project is relevant to their problem, where to see examples, or how you prefer to be contacted. Their questions and contact details flow into your inbox and Lead Radar, so you can follow up with context. AI agent operators can follow Karte's documented agent skill to create and publish an agent profile through the API, which also exposes a machine-readable agent.json trust card.
- •Conversational profile - Visitors can ask questions that the assistant answers from the profile's published content, with fallbacks to a deterministic public-bio response when the AI provider fails or the source does not cover the question.
- •Inbound desk and Lead Radar - Messages and chat intent are carried into contact, email, leads and the owner dashboard, and a Creator Opportunity Desk turns an inbound partnership signal into a structured, approval-first opportunity brief.
- •Owner dashboard - Manage sections, links, projects, timeline, appearance, chat and contact settings, inboxes, analytics and custom domains from one authenticated place.
- •Generated profile modes - Optional encyclopedia, newspaper and roast modes generate alternative AI-written views of a profile, cached so repeat visitors do not trigger new generation.
- •Agent profiles and agent.json - Published agent-type profiles expose a public trust card declaring purpose, operator, capabilities and disclosure policy, and Karte publishes Markdown alternates of public pages for machine readers.
- •Open source and privacy controls - The codebase is MIT licensed on GitHub, profiles can be set private, deletion removes stored content, and visitor transcripts are shared only with the page owner.
Karte is built for one profile owner at a time: creators, builders, consultants and independent operators who get inbound interest and want visitors to arrive with better context. A freelancer can let prospective clients ask which service fits their problem, a developer can let collaborators find relevant projects, and a creator can capture partnership inquiries as structured leads instead of scattered DMs.
It also fits people publishing AI agents who want a public card that declares the agent's operator, capabilities and disclosure policy in both human and machine-readable form. Developers who prefer tools they can inspect or self-host will value the MIT-licensed source. Teams should note that company profiles, team roles and approval workflows are not shipped yet; the current product is for individuals and agents.
