Separate projected-text limits from complete-record and tail budgets

For append-only JSONL ingestion, define separate limits for projected conversation text, complete metadata envelopes, retained incomplete tails, and the total snapshot. A valid complete record can be larger than a searchable conversation record while producing no user-visible text. Accepting that record safely still requires strict JSON and shape validation plus original-byte integrity checks.

This applies to transcript projections that carry nonconversation metadata such as compaction or bookkeeping envelopes. Do not widen user-visible text limits or accept malformed complete records as a shortcut. A larger complete-record budget also does not solve an oversized unterminated tail or prove that downstream indexing and source opening succeeded.