Why does the AI Assistant say the chat is too long?

The TeleportHQ AI Assistant sends your whole conversation with every message, up to a memory limit called Max context. When a chat fills that limit, a Context limit reached banner appears and sending pauses. To continue, choose where the assistant's memory should start, summarize the chat into a new one, raise Max context, or start a new chat with New.

What you see#

• At 85% of the limit: a line above the message box says Context at {n}% of the set limit, with a Manage context link.
• At 100%: the banner Context limit reached says "The conversation is using {n}% of the currently set Max context limit, so sending is paused. Pick where the assistant’s memory should start to continue." The message box reads Context limit reached — pick where the assistant’s memory starts above.
• If you send anyway, the chat shows: "The conversation no longer fits the assistant's context limit. Pick where its memory should start (above the message box) to continue."

The banner also has What does each option do?, which explains the choices below.

How to fix it#

Pick one of these options in the TeleportHQ AI Assistant's Context limit reached banner.

Option 1: Start the memory from a later message#

Use this when the early part of the chat no longer matters, for example after you finished the hero and moved on to the footer.

1. In the banner, open the Start from menu. The recommended message is marked recommended.
2. Pick the message where the assistant's memory should start.
3. Click Set start. A Context starts here divider appears in the chat, and you can send messages again.

Earlier messages stay visible in the chat; they're no longer sent with new messages. You can also set the start point before you hit the limit: hover a message you sent and click the anchor icon (Set anchor — the assistant's memory starts at this message). Click it again to clear it.

Option 2: Summarize into a new chat#

Use this when you want the assistant to keep the gist of the whole conversation.

1. In the banner, click Summarize into a new chat.
2. The AI writes a summary, archives the full chat, and continues in a new chat that starts from that summary.

The archived chat stays in your chat history. Summarizing uses AI tokens.

Option 3: Raise Max context#

Use this when the assistant needs to remember more of this chat.

1. Open Max context in one of these places:
   • in the banner;
   • in the AI Assistant's Settings, under Engine;
   • in Project settings → AI → Assistant, under Memory.
2. Choose a higher value: 200k, 300k, 400k (default), 500k or 600k characters.

Option 4: Start a new chat#

If your task is done, click New at the top of the AI Assistant and continue in a fresh chat. This is usually the cheapest option. See How to edit your website by chatting with the AI Assistant.

Why chats fill up#

• Every message you send and every reply from the TeleportHQ AI Assistant counts toward the limit.
• Each attached image counts as 6,400 characters.
• Long tasks with many steps, such as building a database and then several pages, fill a chat faster.

Nothing is dropped from the chat automatically. The assistant waits until you choose.

Other messages you might see#

"The anchor can’t move — there’s no later message to start from." Your last message already starts the memory. Summarize the chat, fork it from a message, or raise Max context.

"Could not summarize the conversation. Please try again." Click Summarize into a new chat again. If it keeps failing, start a new chat with New.

"It's been a while — your next message will use extra tokens to re-process and cache the conversation. Start a new one if current task is complete." This isn't a limit. It shows when a chat longer than about 40,000 characters has had no message for 5 minutes: the next message costs more. Start a new chat if you're done with the task.

Good to know#

• Keep one chat per task. Short chats remember what matters and cost less per message.
• For instructions the AI should always follow, such as "Always use British English", don't repeat them in a long chat. Add them once under Custom instructions in Project settings → AI → Assistant. See How to give the AI your business details and instructions (Project brief).
• For token costs, see What uses AI tokens, and how much?.