Quasar error messages in the AI panel now use direct, conversational copy — clearer gateway, rate-limit, and context-length guidance without jargon or em dashes.
When Quasar hits a limit or the gateway rejects a request, the chat panel now shows shorter, plain-language errors — easier to scan and act on, in English and four more locales.
What we shipped
The Quasar composer in the site editor shows inline errors when a request fails — no active site, credits exhausted, gateway auth rejected, rate limits, or a chat that outgrew the model's context window.Those messages were rewritten for plain voice:
No em dashes — shorter sentences instead of stacked clauses
Direct instructions — "Wait a few seconds and try again" instead of abstract provider jargon
Clearer gateway troubleshooting — API key rotation, .env.local formatting, and cold-restart steps spelled out as a checklist
Simpler context-length guidance — "This chat is too long for the model" with a clear next step: start a new chat or pick a larger model
The same copy ships in English, German, Spanish, Hungarian, and Polish so localized editors see the same tone.This builds on the clearer AI credit messages from earlier this week — org monthly credits vs platform provider limits stay distinct; this pass tightens wording across every error type.
Why it matters
Error copy is part of the product surface. When a message reads like a log line, teams stall — especially on gateway auth or rate-limit failures that need a quick retry, not a support ticket.Plain voice keeps the focus on what to do next without changing any behavior under the hood.
Recommended workflow
Read the first line — it states the situation in one sentence.
Follow the checklist for gateway auth errors (key rotation, .env.local format, dev-server restart).
Distinguish credit types — org allowance vs platform provider outage; see Chat and suggestions for the full table.
Start a new chat when context-length errors appear after a long session.
What is next
We continue tuning Quasar error copy as models and gateway plans evolve. For the full list of limits and recovery steps, see Chat and suggestions and AI overview in platform docs.
