AI chat removed: bring your own AI tool
The built-in AI chat assistant is gone. The reason is simple: you already use your own AI tool. Here's what stays, what's gone, and what you need to do.

Note (2026-09-03): The coding runners listed here under "What stays" have since been removed from Spedy as well. Spedy no longer runs agents — your own agent (Claude Code, Codex, Cursor, …) works with Spedy directly via MCP. Background in Runners, live previews and agent skills removed. Everything else in this post still applies.
We've removed the built-in AI chat assistant from Spedy. The reason is simple: you already have your own AI tool open anyway — in the browser, in your editor, on your phone. A second chat window inside Spedy didn't add anything. Instead of maintaining yet another chatbot, we're focusing on the AI that actually moves your projects forward — and leaving the conversation to the tool you already trust.
What's gone
The chat assistant has disappeared everywhere it used to show up:
- The chat page and the AI assistant button in the app are gone.
- The Create ticket dialog no longer has a chat mode. You create tickets the normal way through the form — title, description, priority, type.
- The Account → AI page with your personal AI settings is gone.
- The one-time AI consent step is gone entirely — there's nothing left to confirm before an AI feature runs.
If you never actively used the chat, your day-to-day barely changes. If you did: just open your usual AI tool alongside Spedy and keep working as before.
Less to manage in billing and settings
With the chat goes all the management overhead around it:
- No more AI token allowance. Under Settings → Billing there's no token meter and no monthly AI token limit to keep an eye on.
- No AI consent toggle in your organization settings.
The result: fewer switches, less explaining to do for your team, and a billing page that only shows what actually matters — plan, seats, and storage.
What stays
The AI that sets Spedy apart from a plain ticket tool stays exactly where it is:
- AI knowledge base: semantic search across your team's Q&A — questions, answers, solutions, and patterns, cleanly versioned — remains part of the Pro plan.
- Coding runners: agents that work directly on tickets keep running on your own Anthropic key. The AI suggests, your team reviews and approves. You stay in control, with full cost transparency.
One small, visible change in the Knowledge Hub: the Consolidate with AI option is gone. Consolidation is now a manual step — you write the merged version yourself and decide what ends up in the entry. In the same spirit: for the AI that Spedy operates, no project content leaves the platform to an external AI provider. The knowledge base runs on self-hosted embeddings; the coding runners continue to use your own Anthropic key.
What you need to do
Nothing. The chat page and the AI settings disappear automatically once the update is live. Your tickets, projects, comments, and knowledge entries are unchanged.
Two notes:
- Past chat conversations are not retained. If you need something important out of a chat, save it first — as a ticket or a knowledge entry, for example.
- Bookmarks to
/chator/account/aiwill no longer resolve. Clean them up when you get a chance.
Less surface, clearer ownership: you bring the assistant you trust — Spedy takes care of your projects, your knowledge, and your agents.
Frequently asked questions
Quick answers to the most common questions about this topic.
Why was the AI chat removed?
Can I still use AI in Spedy?
Do I need to change anything?
What happens to my AI token allowance?
Keep reading

Runners, live previews and agent skills removed: Spedy is the layer your agents work in
We no longer run agents. What's gone (runners, previews, skills), what stays (MCP, handoff, budgets, approvals) and what you need to do: nothing.

Capacity contracts: clients buy points, not hours
Clients buy a quarterly point allowance, and every request gets a size class before work starts. The price is fixed before anyone picks up the ticket.

Budgets now count billable hours – and weekly hours are finally editable
Budgets measure only what gets charged. Plus: weekly hours per person, approving agent time on the ticket, and credit notes in the invoice list.