Runtime packaging uses an exact-pinned official slim base and a multi-stage,
bundled entrypoint to reduce container loading costs in short runs.
Enforce documented batch and Unicode length limits.
Redact transport/internal error diagnostics and reject mismatched query payloads.
Preserve full request/failure/byte accounting at row limits and bound source retries by the run deadline.
Report unsupported platform resources separately while preserving the input.
Refresh pinned transitive dependencies and Store copy/schema help without changing price or output fields.
Verify new-form prefill, repeat and default resolver with bounded Cloud evidence.
0.1.4 - Public release and post-Publish closeout
Uploaded the release candidate to Apify as latest
Added PPE-aware Dataset writes for the suggestion-saved
Reduced public hard caps to 500 generated Google requests and 5,000 saved
suggestions.
Cleaned customer-facing README, Store listing draft and input help text.
Validated default, capacity and real PPE charge-limit runs before public
publication.
Published the exact 0.1.4/latest
Passed one bounded public default canary and reconciled its Dataset, run
summary, PPE event count and platform usage.
0.1.1 - Technical acceptance
Hardened local npm start
Added a deterministic regression for stable row order when concurrent
responses finish out of order.
Recorded independent technical acceptance evidence for private build 0.1.3
Implemented the first private MVP for Google Autocomplete Scraper.
Added direct HTTP source client, expansion planner, deterministic IDs, dedupe,
Dataset output and RUN_SUMMARY
Added deterministic tests, TypeScript build, Dockerfile and Apify schemas.
Added private MVP validation and handoff documents.
Prepared documentation-only repository scaffold for audit.
Added draft Apify metadata/schema placeholders.
Added workflow, playbook context, local skills and evidence boundaries.
