Home Troubleshooting & FAQ

Troubleshooting & FAQ

By George Pu
• 6 articles

How do I fix common Vinci problems?

How do I fix common Vinci problems? When a message fails to send or Vinci pauses mid-response, one of a few things is happening. Use the error message to choose the right fix. You have run out of credits If Vinci shows "You've reached your limit" or HTTP 402 with budget_exhausted, the shared Vinci balance is out of credits. 1. Open Account in the app or visit the Usage page on the Vinci Platform. 2. Check your credit balance and usage. 3. Add credits, with a $10 minimum purchase, or enable auto top-up. 4. Try your message again. The same balance is used across the web app, mobile apps, API, and Vinci Code, so activity on any of those surfaces can affect what remains. To investigate unexpected usage, review the 7-, 30-, or 90-day view, choose a custom date range, or export a CSV. Vinci is briefly busy If Vinci says "Try again in a moment," reports rate_limit_error, or returns HTTP 429, wait a few seconds and retry. Adding credits does not resolve a rate-limit response. Your API key is the problem If an API request returns HTTP 401 with authentication_error: 1. Confirm that you copied the whole key. It should start with vinci_live_. 2. Check that the key has not been revoked. If it was revoked or lost, issue a new one on the Vinci Platform. 3. Use Authorization: Bearer YOUR_KEY in the request header. Other quick fixes - Cannot sign in: use the same sign-in method you registered with, try Reset password, and check spam or junk for the verification email. A private browser window can rule out an interfering extension. - A file will not attach: each message accepts up to 5 files, with an 8 MB maximum per file. Split or trim a larger file. For more fixes, read the Troubleshooting guide. If the response still does not make sense, email support@getsimpledirect.com or use the live-chat widget on the web app's /support page. Include your account email, whether the issue is in the app or API, what you tried, and the exact error. Do not send a full API key.

Last updated on Jul 24, 2026

My API key isn't working — what do I see?

If your API calls are failing, the error code tells you what's wrong. Here's what each one means and how to fix it. 401 — Authentication error The error: Your key is wrong, expired, or revoked. How to fix: 1. Check that you copied the whole key. It should start with vinci_live_. If it's truncated, copy it again carefully. 2. Confirm the key hasn't been revoked. Go to platform.getsimpledirect.com, open API keys, and check the list. If you revoked it by accident, issue a new one. 3. Check if the key expired. API keys can be set to expire after 30, 90, or 365 days — if yours has expired, issue a new one. 4. Verify the auth header format. Use Authorization: Bearer YOUR_KEY_HERE in your request headers. See the developer docs for a complete example. 402 — Budget exhausted or insufficient quota The error: Your account is out of credits. How to fix: 1. Open Account in the Vinci app or visit platform.getsimpledirect.com. 2. Check your credit balance. 3. Add credits (minimum $10 purchase) or enable auto top-up to refill automatically. 4. Retry your API call. 403 — Permission error The error: Your key doesn't have permission to call that endpoint. How to fix: 1. Go to platform.getsimpledirect.com and open API keys. 2. Find your key and check its scopes. Common scopes are: - inference — needed to call POST /chat/completions - models — needed to call GET /models - usage — needed to read usage data 3. If your key lacks the scope you need, revoke it and issue a new one with the right scopes selected. (You can't edit scopes on existing keys.) 429 — Rate limit error The error: Vinci is briefly busy and rate-limited your request. How to fix: 1. Wait a few seconds — this is temporary. 2. Retry your request. It should succeed. This isn't a credit or authentication issue — it just means we're handling a surge of traffic. A short retry delay (with exponential backoff) is standard practice. Still getting errors? If none of the above match what you're seeing, contact support with: - The exact error code and message. - The endpoint you're calling (e.g., POST /chat/completions). - A small, reproducible example of your request (don't include your actual key). - What programming language or tool you're using. For more details Check the developer docs for setup instructions and a complete API reference.

Last updated on Jul 24, 2026

Why did my whole chat disappear?

If you started a conversation, left before Vinci finished responding, and then couldn't find it anymore, here's what happened and how to prevent it. What causes this On the web, if you navigate away or refresh the page before a response finishes, Vinci doesn't save the conversation to your account. This happens because: - Vinci sends your message and waits for the response to stream back. - If you refresh, go to another tab, or close the page before the response arrives, the message never makes it to our servers. - The conversation is lost because it was never permanently stored. Mobile doesn't have this issue — your chats are saved safely on your device whether or not they're synced. How to prevent it 1. Wait for responses to finish. Let Vinci complete its reply before navigating away, refreshing, or starting a new chat. 2. If you need to leave mid-response: On the web, use Stop generating (the stop button in the chat) before you leave — this saves what Vinci has sent so far. 3. Turn on device sync if you're worried about losing chats. With sync enabled, conversations are saved to your account even if you navigate away. If it already happened Unfortunately, if the chat was never saved to our servers, it can't be recovered — it only existed on your device in that session, and refreshing cleared it. Going forward: - Migrate to mobile or enable device sync on the web so chats are backed up to your account. - Check your browser's history — sometimes you can recover what you typed in the message box, depending on your browser settings. Vinci on mobile Mobile conversations are safer — they're saved on your device automatically, whether or not sync is on. You won't lose a chat just by navigating or refreshing.

Last updated on Jul 24, 2026

Why does Vinci claim to be Claude or ChatGPT?

If you asked Vinci "what are you?" and it said it was Claude or ChatGPT, you've found a bug. Here's what Vinci actually is and what to do if this happens. What Vinci really is Vinci runs on a model class — a curated set of frontier models we test, tune and rotate, reached through one endpoint. From lightest to deepest, the public classes are Vinci Mezzo, Vinci Forte (the default), and Vinci Fortissimo for the hardest work. Mezzo is fast and light for everyday chat; for code and file tasks, Forte works best. In chat, click the model badge at the top right to switch inline, or choose under Settings → Personalization. Per-app overrides are available. A class is not a single model, and it is not one we trained: we select the strongest available engine, run it on zero-data-retention infrastructure, and swap it as better ones ship. The classes page lists what each class is running right now. (SimpleDirect does build its own open models — Piccolo and Bozza — but those are given away free and are not what answers you here.) So whichever class you are on, you are talking to Vinci — not Claude, ChatGPT, Gemini, or any other product. It should always present itself that way. When it gets confused Occasionally, Vinci will claim to be a different model — usually Claude or ChatGPT. This is a known issue in beta and happens rarely, but when it does, it's a bug worth reporting. If this happens to you: 1. Screenshot or note the exact response where Vinci misidentifies itself. 2. Contact support with: - The exact question you asked and Vinci's exact response. - Which platform (web, mobile app, or API). - The date and time it happened. 3. Include the screenshot if you have one. Why it matters We want you to know what you're actually using, and identity confusion breaks that trust. Reporting these incidents helps us fix the issue faster. For developers If you're building on the Vinci API, model takes a class id — mezzo, forte, or fortissimo — or auto to follow the account's choice. Omit it entirely and you get auto, which resolves to the account preference (Forte by default). piano remains reserved and unavailable. See the developer docs for full API details.

Last updated on Aug 03, 2026

Why is Vinci slow to respond?

If you've noticed that Vinci takes a few seconds to start replying, you're noticing something real. Here's what's happening and whether it's a problem. A few seconds before the first word Vinci typically takes 1–2 seconds before the first word appears on screen. This is normal and not a bug — it's the time needed to: - Send your message to the model. - Process it. - Start streaming the response back to you. This delay is baked into how the service works and affects everyone equally. You'll see this every time you send a message, and it's not something you can speed up on your end. When slowness is a real issue If you're seeing delays longer than 5–10 seconds, or if responses stop mid-stream, something else might be happening: - Check your internet connection. A slow or unstable connection can make responses take longer. - Try a different browser or device. Browser extensions can sometimes interfere. - Refresh the page on the web app to clear any stuck connections. The model is also thinking Depending on your question, Vinci might also pause after you see the first words while it thinks through a harder problem. This is intentional — more complex questions get more thinking time. Still feeling slow? If responses are consistently taking a very long time or timing out, contact support with details about: - What device and browser you're using. - How long responses typically take (with an example of a simple question). - Whether it's happening in the app or through the API. A few seconds at the start is expected, but consistent slowness beyond that is worth investigating.

Last updated on Jul 24, 2026