Don't stop here
Hand-picked guides our readers explore right after this one.
Expert guide to Claude prompts with XML tags, artifacts, and complex reasoning
Read the guideStunning image generation with Midjourney prompt mastery
Read the guideData analysis workflows with prompt engineering
Read the guide'Your message will exceed the length limit for this chat' and 'this conversation is too long' are Claude telling you it is at or near its context window. It is a feature, not a silent failure: unlike tools that just truncate, Claude warns you before it loses the thread. Two things cause it, a single message (often a big paste or file) that is too large on its own, and a conversation that has simply grown too long across many turns. The fix is the same idea either way: reduce what you are sending, or start a fresh chat carrying a summary so you keep the context without the bloat. On paid plans with code execution, Claude auto-summarizes earlier messages as it nears the limit, so most users hit this rarely. This guide covers both the single-message and long-conversation versions.
'Your message will exceed the length limit for this chat' appears before you send
'This conversation is too long' or similar blocks further replies
A large paste or file upload triggers it immediately
It appears deep into a long working session and stops progress
You are told to click Continue, then told continuing is not possible
One message, often a long paste, a big document, or a large code file, exceeds the maximum input for the chat on its own. Claude flags it before sending rather than failing silently.
Every turn stays in context, so a long back-and-forth eventually fills the window. Once full, Claude cannot take more without losing the earliest parts, so it warns you.
Uploaded files count toward the context. Several big files, or one very large one, can eat most of the window and leave little room for the conversation.
Some builds first suggest clicking Continue, then state continuing is impossible, and starting a new chat loses context, a reported bug that makes the limit feel worse than it is.
When to try: First, when a big paste or file triggers it
If one message is too big, split it and send the parts in sequence, for example 'Part 1 of 3, do not respond yet', then continue. Claude assembles the context across the parts without hitting the single-message cap.
When to try: When a long conversation is the cause
Before you hit the wall, ask: 'Summarize everything important from this conversation, what we are doing, key decisions, current state, in about 400 words.' Copy that into a fresh chat as the opening context. You keep continuity without the accumulated bloat.
When to try: When large source material is filling the window
Instead of a whole document, paste just the section Claude needs, or upload the file in a separate, dedicated conversation and pull only the relevant parts. Less input means more room for the actual work.
When to try: As soon as you see the warning
Claude shows the warning before the message goes through, so trim the message or start a new chat first. Acting on the warning avoids the harder mid-task version where you are already blocked.
When to try: For recurring work with the same reference material
For ongoing work, put stable reference material in a Claude Project's knowledge instead of pasting it into every chat. The Project holds context outside the single conversation, so individual chats stay lighter.
When to try: For big multi-session projects
Rather than one endless thread, run the work as a series of shorter chats, each handed the summary from the last. This is the durable pattern for long projects and avoids the limit entirely.
Carry a short summary into a fresh chat instead of running one endless thread
Paste only the excerpt you need, not entire documents
Put stable reference material in a Project rather than re-pasting it
Act on the length warning when it first appears, before you are blocked
This is expected behavior, not a fault, so support cannot raise the per-chat limit. Contact Anthropic via support.claude.com only if you hit contradictory 'continue then cannot continue' prompts or the warning appears on clearly small messages, which points to the known bug rather than a genuinely full window.
Claude is at or near its context window for that conversation. Either the single message is too big, or the chat has grown too long overall. It warns you before sending rather than silently dropping context. Trim the message or start a fresh chat with a summary.
Before restarting, ask Claude to summarize the conversation, goals, key decisions, and current state in about 400 words. Paste that summary as the first message of a new chat. You keep continuity without carrying the accumulated length.
No, it is fixed by the model's context window and no plan raises the per-conversation cap. On paid plans with code execution, Claude auto-summarizes earlier messages as it approaches the limit, but the ceiling is the same. Manage it by chunking and summarizing.
That is a known 2026 issue where the prompts contradict each other and starting a new chat loses context. Work around it by proactively summarizing and moving to a fresh chat before you hit the hard wall.
Yes, attachments count toward the context window. Several large files, or one very big one, can fill most of the window. Upload files in a dedicated chat and pull only the parts you need, or keep reference material in a Project.