Where do I find the official list of Base44 error messages?

There isn't a complete one. Base44's documentation has a Troubleshooting section at docs.base44.com that covers the most common visible errors, but it does not enumerate every error code or symptom. The platform's error messages are also inconsistent — the same underlying issue may surface as different messages depending on where in the request flow it triggered. This reference fills the gap by aggregating documented platform errors with the ones we see most often in client engagements.

What error means my backend function failed to deploy?

Most often ISOLATE_INTERNAL_FAILURE, sometimes accompanied by 'Backend function URL error' in the browser. The cause is almost always a syntax error or an unsupported import in the function code. Open the function in the IDE, check for typos, and verify all imports are supported by the Deno runtime. If the error persists with clean syntax, the platform may have changed its supported import list — check status.base44.com.

Why am I getting 405 Method Not Allowed on my POST endpoint?

A documented routing bug. Base44's request router occasionally treats /functions/* paths as frontend routes rather than backend, returning 405 instead of forwarding to your Deno handler. Workaround: prefix the function path with a known-routed segment, or use a backend function pattern that does not collide with frontend routes. We have the full diagnosis in [the function routing fix](/fix/backend-functions-404-routing-broken).

What does 'Token expired' mean in a Base44 backend function?

The user's JWT exceeded its TTL during the function's execution. By default, Base44 tokens expire after a few hours. If a long-running operation calls the SDK after the token expires, the SDK throws a 401. Fix: catch the 401 in the function, return an explicit 'session expired' response, and have the frontend re-authenticate. Don't try to refresh the token inside the function; there is no documented service-account mode.

Why did Base44 charge me for a generation that failed?

The platform charges for AI inference regardless of whether the result was usable. Sandbox timeouts, unsupported dependency rejections, and partial generations all cost credits. There is no automatic refund mechanism. The pattern that helps: scope prompts narrowly so failures cost fewer credits, and snapshot before each prompt so reverting is cheap. We track this issue in [the credit burn fix article](/fix/excessive-credit-burn-minor-changes).

What's the fastest way to diagnose a 500 error from a Base44 backend function?

Check the platform's function logs first; they show stack traces for unhandled exceptions. If the logs are empty or unhelpful, add explicit logging to your function (console.log with a prefix that's easy to grep), redeploy, reproduce, and inspect. For production diagnosis, ship logs out via fetch to an external service so you can grep historical errors. Without external logging, post-incident forensics is difficult.

Base44 Error Reference: 30+ Documented Errors and Their Fixes

Why this matters

Most Base44 error troubleshooting is fragmentary. The platform's docs cover some errors. Feedback boards have user-reported errors with mixed-quality answers. Stack Overflow has a few. Nowhere is there a complete reference. This article aggregates what we have observed across audit engagements with one-line fixes per error.

The fixes here are surface-level. For errors that need deeper diagnosis — regression loops, performance problems, security issues — the row links to the relevant /fix/ page where we go deep. Use this reference as the first lookup; escalate to the linked page when the one-liner doesn't solve it.

How to use this reference

Find your error in the table by symptom or message.
Read the cause column to confirm the diagnosis matches your case.
Apply the fix.
If it doesn't work, follow the /fix/ link for deeper guidance.

We update this list quarterly. If you hit an error not in the list, contact us and we'll add it.

Generation and build errors

Error	Symptom	Cause	Fix
Sandbox timeout during AI build	Generation hangs, then errors with "sandbox timeout"	Single prompt too large for the model's context	Split the prompt into smaller scoped requests, one component at a time
Unsupported dependency rejection	"Package X is not supported" during build	Backend function imports a package not in Deno's curated list	Replace with a supported package or call the dependency over HTTP from the function
Build fails — no clear reason	Deploy fails with generic error	Most often a syntax error in agent-generated code; sometimes a platform-side issue	Check the IDE for red underlines; if clean, check status.base44.com
npm package rejected	Specific package fails import	Package uses Node-only APIs (fs, child_process)	Use a Deno-compatible alternative or move logic to client side
Generation timeout	Agent stops responding mid-generation	Context window exceeded or platform-side latency spike	Refresh chat, scope the next prompt narrower, snapshot before retrying
Hallucinated field references	Code references a field that doesn't exist on the entity	Agent forgot the actual schema due to context loss	Re-prime the chat with a paste of the actual schema; see hallucinated fields fix

Deployment and routing errors

Error	Symptom	Cause	Fix
ISOLATE_INTERNAL_FAILURE	Backend function returns 500 with this message	Deno couldn't load the function — usually a syntax error or unsupported import	Open the function in IDE, fix syntax, verify imports are Deno-compatible
Backend function URL error in browser	"Network error" when calling functions.someFunction	Function hasn't deployed or the URL routing is broken	Verify the function shows as deployed in IDE; if yes, see routing fix below
405 Method Not Allowed on POST	Function accepts POST in code but returns 405	Routing bug where /functions/* gets caught by frontend router	Use the workaround in the linked fix article
Deployment failed with no clear reason	Publish completes but new code isn't live	Platform deployment partial-failure	Re-publish after a few minutes; if persistent, file ticket
Published changes not appearing	New code in IDE, old code in production	Browser cache or platform-side caching	Hard refresh (Cmd+Shift+R), then check via incognito to confirm
Domain stuck in pending	Custom domain shows "pending DNS verification" indefinitely	DNS propagation delay or wrong record type	Verify DNS via dig/nslookup; full guide in domain DNS pending fix
ERR_SSL_PROTOCOL_ERROR	Domain doesn't load over HTTPS	SSL provisioning hasn't completed or DNS is wrong	Wait 15–30 min; if persistent, re-add the domain
429 on domain reconnect	Can't re-add a previously-deleted domain	Rate limited at platform level	Wait 24 hours, then retry

Runtime errors

Error	Symptom	Cause	Fix
White screen on click / page unresponsive	Page loads, then goes blank on interaction	Unhandled JS exception in the React tree	Open browser DevTools console; fix the error; redeploy
White screen for authenticated users only	Logged-out users see content, logged-in see blank	Auth-dependent code throwing on a missing field	Check the User entity for missing fields; add defaults
Blank screen after publishing	App was working, now blank	Field mismatch — UI references a field that no longer exists on entity	Restore the missing field or update UI to match new schema
App crashes on mobile	Works on desktop, crashes on phone	Memory or unsupported-API issue	Test in mobile Safari; check for desktop-only APIs (window.fs, etc.)
Buttons stop working after hours	Features that worked become non-functional	Often token expiry or platform-side cache invalidation	Hard refresh; if persists, re-deploy and check Token expiry handling
Editor hangs / extended loading	Base44 IDE itself becomes unresponsive	Project size has grown beyond efficient editor performance	Refactor large files; close other browser tabs; clear IDE cache
Apps fail after revert	"Unexpected error: contact support@modal.com"	Revert left runtime in inconsistent state	Re-publish current state; if needed, re-deploy from a known-good snapshot
'dict' object has no attribute 'lower'	Chat error during a prompt	Platform-side issue with prompt parsing	Refresh chat, retry the prompt with simpler phrasing
Chat stuck on Thinking...	First prompt of a session hangs	Cold-start latency or platform-side queue	Wait 60s; if no response, refresh the chat
Discuss mode uses too many credits	Credits drop fast even without code changes	Each discuss-mode turn consumes credits	Switch to direct edit mode for routine work

Data and entity errors

Error	Symptom	Cause	Fix
Data disappears on return to app	Saved data vanishes when user logs back in	Often an Entity.list filter mismatch or schema change	Verify the entity still has the field; check filter; restore from snapshot
Cursor edit loads old version	Code edit overwrites recent changes	Editor sync issue with stale state	Refresh; manually merge via version history
5,000 item limit exceeded	List call truncated at 5,000 records	Platform limit since November 2025	Paginate with cursor on a sortable field
RLS not restricting data	Users see records they shouldn't	RLS rules misconfigured or not applied	Audit RLS rules; test with second account; full guide in security checklist
Workspace move irreversible	App now in wrong workspace, can't move back	Documented platform limitation	No fix; recreate the app or contact support
Database schema mismatch	Code references field, error says field doesn't exist	Schema drift between code and entity definition	Update entity to match code, or vice versa
File upload fails — invalid MIME type	File rejected even with valid format	Browser-reported MIME type doesn't match expected	Re-encode the file or relax MIME validation server-side
File upload exceeds 50MB	Large file upload fails	Platform per-file limit	Compress, split, or use external storage (S3) for large files

Authentication errors

Error	Symptom	Cause	Fix
Email verification loop	User can't get past verification step	Verification email lands in spam or platform-side issue	Check spam; resend; if persistent, contact support
Google OAuth redirects to error	Google login fails after consent	OAuth callback misconfigured or domain mismatch	Verify allowed origins/redirects in both Google Cloud and Base44
Forgot password no email	Reset email never arrives	Often deliverability via shared sender pool	Use external email provider; check spam; verify email exists
Token expired (401 in function)	Backend SDK calls return 401	User's JWT expired during long-running op	Catch 401, return session-expired response, force re-auth on frontend
User locked out	Can't log in despite correct credentials	Rate limit on failed auth or account flag	Wait 15 min; check Base44 admin panel for account state
SSO bypass risk (pre-July 2025)	Anyone can register on private app	The patched-but-structural issue	Verify org-domain check post-signup in a backend function

Integration errors

Error	Symptom	Cause	Fix
Stripe payment not processing	Checkout fails or succeeds but no record	API key mismatch (test vs live) or webhook not signed	Verify keys; validate webhook signatures; full guide in Stripe integration guide
Webhook only fires when users active	Subscription renewals fail at 3am	Documented platform behavior	External scheduler (cron-job.org) triggers a backend function
Webhook returns 500 on POST	Function deployed but webhook fails	Wrong URL, missing signature validation, or routing bug	Verify URL; check function logs; see routing fix
Stripe in-app purchase rejection	iOS app rejected for missing StoreKit	Apple requires StoreKit for digital goods	Use native iOS shell with StoreKit; route subscriptions through it
Hugging Face 401 from Base44 app	API returns 401 with valid key	Often token leakage or platform-side proxy issue	Move HF calls through backend function with secret in env vars
Custom integration deprecated	New integration won't add	Custom integrations deprecated post-March 2026	Use a backend function calling the third-party API directly
Twilio SMS not sending	Send call succeeds but no message	Twilio account state, invalid number, or rate limit	Verify Twilio dashboard; check Twilio logs

Billing and credit errors

Error	Symptom	Cause	Fix
Excessive credit burn for minor changes	Small edits cost 10+ credits	Agent regenerates large code regions	Scope prompts narrowly; full guide in credit burn fix
Cannot purchase additional credits	UI shows option but purchase fails	Platform requires tier upgrade for more credits	Upgrade tier or wait for next cycle
Unused credits don't roll over	Remaining credits gone at month-end	Documented platform behavior — no rollover	Plan usage to consume the allowance each cycle
Hidden dual subscriptions	Charged twice per cycle	Platform billing bug — UI shows only one	Contact billing support; document both invoices
Pricing increased post-Wix	Tier costs more than at signup	Acquisition-era pricing changes	Re-evaluate tier fit; budget 25% headroom
Monthly plan not unlimited despite name	Hits credit cap on Monthly tier	Platform tier mapping not what naming suggests	Read tier docs carefully; choose tier on credit allowance, not name

Performance and rate limit errors

Error	Symptom	Cause	Fix
429 Too Many Requests	Sudden burst of failed API calls	Platform rate limit on burst traffic	Add exponential backoff; reduce concurrency
500 Internal Server Error from API	Sporadic failures on entity calls	Platform-side issue	Retry with backoff; check status.base44.com
Scheduled task not running	Cron-style job doesn't fire	Platform's scheduled functions are unreliable when no users active	External scheduler triggers the function via fetch
Editor extremely slow	IDE laggy on every keystroke	Project size or browser memory	Refactor; close tabs; clear IDE cache

Mobile and app store errors

Error	Symptom	Cause	Fix
App Store rejection — web wrapper	iOS submission rejected	Apple's "essentially a web view" policy	Add native features beyond the web view; or pivot to PWA-only
Build number not incrementing	Same IPA file rejected by App Store	Build version not updating between exports	Manually bump build number in export config
Push notifications not supported	Mobile users don't get notifications	Platform limitation on managed mobile builds	Use a third-party push provider (OneSignal) wired via custom code
Android camera missing	Camera UI doesn't appear on Android	Manifest permission missing	Add camera permission to Android manifest in build config

Platform-wide and outage errors

Error	Symptom	Cause	Fix
Platform outage	Multiple apps down simultaneously	Base44 infrastructure issue (Feb 3 2026 example)	Check status.base44.com; wait; no SLA recourse
Support ticket no response	Days/weeks with no reply	Documented post-acquisition support degradation	Escalate via public channels (Twitter, feedback board) if critical
Status page misleading	App down, status page green	Status page only reflects platform-wide outages	Maintain external synthetic monitor as ground truth

Common error-handling mistakes

Treating every error as transient. Some errors require code fixes; retrying without diagnosis just burns credits.

Ignoring the function logs. They are the first place to look for backend errors and are routinely skipped.

Not shipping logs externally. Platform logs roll quickly. Without external retention, post-incident debugging is impossible.

Using shared error messages. Wrap every SDK call in try/catch and surface user-friendly messages. Default platform errors leak implementation details.

Skipping snapshots before risky operations. Many errors are recoverable via snapshot revert. No snapshot, no recovery.

Want us to triage your Base44 error log?

Our $497 audit reviews your function logs, identifies the top error patterns, and produces a prioritized fix list. Most apps have 5–15 distinct recurring error types; we resolve the top 3–5 in a 48-hour fix sprint after the audit. Order an audit or book a free 15-minute call.

Base44 SDK Reference — the API surface and its specific failure modes.
Base44 Credit System Explained — billing-error diagnosis and prevention.
Base44 Production Readiness Guide — the broader operational picture, including observability that catches errors early.