Ownership and preparation: GPTMarketPlus and Agent Foundry share ownership with the products described here. This commercial introduction was prepared with AI assistance. An automated development agent ran the included fictional examples on November 8, 2026; the shared editorial byline does not imply independent testing or a named human review.
During an incident, billing notes, dashboard states, application errors, and support replies can leave you with a scattered record. The Supabase Recovery Evidence Agent organizes already-redacted observations into a local incident packet. It costs $1.00 USD through PayPal, or 1.00 USDC on Base, as a one-time toolkit purchase.
This is an independent product and is not affiliated with Supabase. Its role is evidence organization. It does not log into an account, restore a database, change DNS, or send anything to support.
Begin with the fictional incident
The ZIP includes Python source, a sample input file, and tests. It requires Python 3.10 or newer and no third-party Python packages. Extract the archive, open a terminal in the folder containing app.py, and run:
python3 app.py examples/sample_incident.json --output-dir output --clock 2026-08-29T01:00:00ZUse your installed Python launcher if it is named python or py -3, after checking its version. The fixed clock belongs to this fictional example and makes it repeatable; it is not the current time or a suggested timestamp for your own incident.
Our automated sample run used Python 3.13.9, exited with status 0, and created output/report.json and output/escalation.md. Open both files locally to understand the packet before adapting the sample input.
Prepare a record you can safely review
Follow the included input example and README. Keep observations tied to their source identifiers and original timestamps. Record what you actually observed, including uncertainty and missing details, instead of filling gaps with guesses. Billing state, control-plane state, service health, and DNS observations describe different parts of an incident.
Remove passwords, API keys, access tokens, personal information, and other sensitive material before creating the input file. The toolkit's secret-pattern checks are an aid; they do not make arbitrary raw logs safe or replace redaction. Use only information you are authorized to process.
Review the packet before sharing it
The toolkit is designed to produce a timeline, documented contradiction findings, missing-evidence checks, and an escalation draft. Read each finding against its supplied source. A mismatch or missing item is a reason to investigate; it does not by itself establish what broke or who caused it.
Check the impact statement, timestamps, attempted actions, and support questions in the Markdown draft. Remove anything that is wrong or unsuitable to share, then decide whether and where to send it yourself. The toolkit makes no network calls and does not submit a support ticket.
Get the toolkit
Pay $1.00 USD with PayPal, or check the product page for the USDC option and current scope. The ZIP becomes available after verified payment. For purchase or download help, use purchase support. This purchase provides local evidence tooling; it does not include incident response services or promise a recovery outcome.