“Claude hit the maximum length for this conversation”
Your conversation isn't lost. Paste it below — you'll get a fresh-start message that carries over what matters, so the new chat picks up where this one gave out.
Even easier — paste a share link. On Claude or ChatGPT: open your chat → Share → Copy link, then paste the link here instead of the text. We pull the whole conversation server-side, nothing missed. (A share link makes that chat viewable by anyone with the URL — you can un-share it right after.)
🔒 We never store your conversation. Pasted text is analyzed right in your browser; a share link (and the written rescue) is processed on our server and discarded immediately after — never saved, never trained on.
How this level was decided
Built from your own conversation — the goal, the things that got lost, and where you left off. Edit anything, then paste it as the first message of a new chat.
Why this happens
Claude can only hold a certain amount of conversation in mind at once (its “context window”). When your chat reaches that limit, Claude refuses new messages — that's the error you just saw. It tends to happen right in the middle of real work: long projects, documents, back-and-forth edits.
Doesn't Claude compact conversations automatically now?
Sometimes, yes — Claude may summarize earlier messages to keep going. But an automatic summary decides for you what to keep, and it happens silently: the details it drops are exactly the ones you never get to review. A handoff you can read and edit before sending keeps you in control of what the new chat remembers. That's what the tool above gives you.
Will I lose my work by starting a new chat?
Not if the new chat starts with the right context. The handoff built above quotes your own conversation: your goal, the constraints you had to repeat, and the last state of the work. Paste it as the first message of a fresh conversation and continue from there — fresh chats are also faster and sharper than saturated ones.
How the check works
The page reads your pasted conversation locally and looks for verifiable signs of wear: things you had to repeat, corrections you had to make, answers that came back nearly identical. Every flag it raises is backed by a quote from your own chat — click “why we say this” under any finding. What it can't detect (silent contradictions nobody flagged), it says so instead of guessing.