Chat Interface Troubleshooting
Diagnostic procedures for chat-specific issues including unresponsive sessions, incorrect outputs, missing history, rate limits, truncation, context loss, keyboard shortcuts, and export failures.
Step-by-step fixes for common chat problems. Find your symptom, follow the steps, get back to work.
Quick Diagnosis
Start here. Find your symptom, then jump to the detailed fix.
| Symptom | Most likely cause | First step | Section |
|---|---|---|---|
| No response appears | Model processing delay or service disruption | Wait 60 seconds, then click [Regenerate] | No Response |
| Answer is factually wrong | Hallucination or underspecified prompt | Ask IAN to quote its source | Incorrect Output |
| Chat not visible in sidebar | Unintuitive name | Search History by keyword | Missing Chat |
| "Rate limit exceeded" error | Too many messages in a short window | Wait 5 minutes | Rate Limit |
| Response stops mid-sentence | Model hit output token limit | Type "Continue" |
Truncated Response |
| IAN ignores earlier messages | Conversation exceeded context window | Start new chat with a summary | Context Loss |
| Keyboard shortcuts not working | Message box not focused or browser conflict | Click in message box, press Escape first |
Keyboard Shortcuts |
| Export produces no download | Browser blocking download or chat too long | Check downloads folder, try a different format | Export Issues |
1. No Response
What you see: You sent a message. The loading indicator spins. Nothing appears.
Important: Some questions genuinely take longer, especially with complex models. Wait at least 60 seconds before troubleshooting.
Resolution
Try each step in order. Stop when the issue is resolved.
| Step | Action | Why |
|---|---|---|
| 1 | Click [Regenerate] | Retries your question with the same model |
| 2 | Send a test message: "Hello, are you there?" |
Confirms whether the service itself is responding |
| 3 | Refresh the browser page and retry | Clears stale connection state |
| 4 | Try a different browser (Chrome, Firefox, Edge) | Rules out browser-specific issues |
If none of the above works
The service may be experiencing a temporary outage.
- Wait 10 minutes and try again.
- Check with a colleague to confirm whether the issue is widespread.
- If the problem persists, contact support — include the time, your browser, and what you were asking. Email: gss.sr@xceedance.com.
Network note: If you're on a VPN or corporate network, try disconnecting temporarily. Some network configurations interfere with long-running connections.
2. Incorrect Output
What you see: IAN gave you an answer that doesn't match what you know, or the facts are clearly wrong.
This is called hallucination — the model generates text that sounds confident but isn't accurate. It happens with all AI systems and is a known limitation.
Resolution
| Step | Action | What to look for |
|---|---|---|
| 1 | Ask for a source: "Quote the exact text or source where you found that" |
If IAN can't produce a quote, the answer may be fabricated |
| 2 | Rephrase with more detail | Specific questions get more accurate answers (see examples below) |
| 3 | Click [Regenerate] | The same model may produce a better answer on a second attempt |
Rephrasing examples:
| Instead of… | Try… |
|---|---|
"Is this covered?" |
"Based on the policy document I uploaded, is damage from flooding covered under Section 4?" |
"What are the rules?" |
"What are the reporting deadlines for workers' compensation claims in our jurisdiction?" |
If answers are consistently wrong
- The question may be outside the model's training data.
- The model may not have the information it needs — try uploading a relevant document and asking IAN to answer from that document specifically.
Important: Never rely on IAN as your only source for claims decisions, compliance questions, legal advice, or financial calculations. Always verify with official sources or a human expert.
3. Missing Chat
What you see: You can't find a conversation you had earlier.
Chats are stored in two places:
| Location | What's there | Retention |
|---|---|---|
| History (sidebar) | All recent chats | Indefinite |
| Saved ⭐ (sidebar) | Chats you starred | Indefinite |
Resolution
- Click History in the sidebar.
- Use the search box — type any keyword you remember from the conversation.
- Scroll through recent chats if you can't recall exact words.
- Check Saved in the sidebar — you may have starred it.
If you still can't find it
The chat was likely manually deleted. Deletions are permanent and cannot be undone.
Prevention: Star (⭐) any chat you might need later, as soon as it becomes useful. Starred chats never expire.
4. Rate Limit Exceeded
What you see: An error message about rate limits or too many requests.
Why it happens: You've sent more messages than your allocation allows within a time window.
Resolution
| Step | Action |
|---|---|
| 1 | Wait 5 minutes — most rate limits reset quickly |
| 2 | If this keeps happening, ask your admin to adjust your rate limit |
5. Truncated Response
What you see: IAN's response stops mid-sentence.
Why it happens: The model reached its maximum output length for a single response.
Resolution
| Step | Action |
|---|---|
| 1 | Type "Continue" or "Keep going" — IAN picks up where it stopped |
| 2 | If that doesn't work, ask a more focused question to get a shorter, complete answer |
| 3 | Break complex questions into parts: "First, explain X. Then I'll ask about Y." |
6. Context Loss
What you see: You mention something from earlier in the conversation, and IAN doesn't remember it.
Why it happens: Every model has a context window — the maximum amount of text it can process at once. In long conversations, earlier messages fall outside that window.
Resolution
- Start a new chat.
- In your first message, summarize the key points:
"Here's what we established so far: [summary]. Now let's continue with…" - Continue from there.
Prevention
| Practice | Why it helps |
|---|---|
| Keep conversations focused on one topic | Reduces context consumption |
Periodically summarize progress: "So far we've established X, Y, and Z. Now let's talk about…" |
Reinforces key points within the active window |
| Save and export before starting a new thread | Preserves a record you can reference when summarizing. See Chat Management. |
7. Keyboard Shortcuts Not Working
What you see: Enter doesn't send, Shift+Enter doesn't create a new line, or other shortcuts are unresponsive.
Resolution
| Step | Action |
|---|---|
| 1 | Click inside the message box — shortcuts require the input field to have focus |
| 2 | Press Escape first — a popup or modal may be intercepting input |
| 3 | Check whether a browser extension is capturing keyboard shortcuts (try in incognito/private mode) |
| 4 | Try a different browser |
| 5 | Clear browser cache (Ctrl+Shift+Delete → clear cached files) and reload |
8. Export Not Working
What you see: You click Export but no file downloads.
Resolution
| Step | Action |
|---|---|
| 1 | Check your browser's downloads folder — the file may have downloaded without a visible prompt |
| 2 | Try a different export format (if PDF fails, try Markdown) |
| 3 | Export a shorter chat — very long conversations may cause a timeout |
| 4 | Temporarily disable popup blockers — some browsers block automatic downloads |
| 5 | Try a different browser |
Escalation
If none of the above resolves your issue:
- Note what you were doing and what happened.
- Note what you expected to happen.
- Take a screenshot of any error message.
- Submit via Support & Escalation — include all of the above.
Related Pages
| Page | Purpose |
|---|---|
| Chat Interface | Core chat interaction workflow and feature reference |
| FAQ | Quick answers to frequently asked questions |
| Support & Escalation | How to reach a human for unresolved issues |