SeaTicket

An open-source project's questions rarely arrive in one spot. Someone files a GitHub issue, someone else asks in Discord, a paying customer emails, and the same bug shows up in the forum. SeaTicket syncs those channels into one workspace so public threads and private requests sit side by side. It connects to GitHub, Discord, Discourse and email, and also lists integrations with Notion, Confluence, Linear and Jira.

When a new question comes in, an AI agent searches related issues, past cases and your knowledge base, then suggests a reply or a ticket. A person on the team approves, edits or discards each suggestion before anything goes back to the original thread. If a question needs an engineer, it becomes a ticket with the context attached, and the reply still returns to wherever the question started. Approved answers can be saved as knowledge for the next time. There's also a self-service portal where users can ask an AI chat, browse docs and track tickets.

That approval step is the useful design choice here. The AI does the searching and drafting, and your team keeps control of what's actually said to users.

SeaTicket is Apache 2.0 licensed. You can run it on your own servers with Docker Compose and point its agents at the language and embedding models you choose, including locally hosted ones. The hosted cloud is priced by issue volume and AI credits: free for up to 1,000 issues and 100 credits a month, then $25, $100 and $500 a month tiers, all with unlimited team members.

On the cloud, the privacy policy says customer content sent to AI features may go to outside model providers, naming Google's Gemini and OpenAI's GPT models as examples. Teams that can't send support data to those providers have the self-hosted route. SeaTicket is built and run by Seafile Ltd, which grew it out of supporting Seafile's own users.