Financial Evidence
Privacy
Last updated 7 October 2026.
What the service receives
Financial Evidence requires no account, API key, or payment information. A request contains only the selected research topic or topics and ordinary protocol metadata needed to return a response. Please do not put personal or confidential information into a topic request.
Infrastructure data
The public documentation and remote MCP endpoint are delivered using GitHub Pages, Cloudflare, and the four public product surfaces. Those providers may process standard security and delivery logs such as IP address, user agent, timestamp, requested path, and operational error information under their own policies. The evidence router may make allowlisted upstream requests needed to retrieve the selected public evidence.
Use and sharing
LIQUILENS PRIVATE LIMITED uses operational data only to secure, diagnose, rate-limit, and operate the service. It does not sell personal information, build advertising profiles, or intentionally collect sensitive personal data through this service. Data may be disclosed when required by law or needed to protect the service and its users.
Optional funding download measurement
The funding review page measures downloads only after you opt in. It stores a random browser identifier in your browser and sends it with a fixed download event and packet identifier. The application stores only a keyed hash of that identifier, UTC day, packet identifier and event name. It does not store IP addresses, browser details, names, employers, portfolios or free text in this counter database. Events describe a download initiated after integrity checks; they do not prove saved files, fund identity or research use.
Measurement starts off. Uncheck the option or choose “Forget my usage” to disable it and request removal. If deletion is unavailable, the page keeps measurement disabled and lets you retry. Counters expire on daily maintenance within 30 days, including when no new usage arrives. Browser identifiers expire after 30 days. These counters and their key are excluded from evidence backups. Python clients send no browser usage events. Standard infrastructure delivery logs remain separate.
Optional application measurement
The funding automation API works anonymously. Its anonymous request counts contain only UTC day, traffic class and outcome. Creating an optional application key explicitly opts into server-side measurement linked to a random installation identifier. The service stores a hash of the high-entropy key, registration and last-use dates, classification, first-use dates and deduplicated packet completions. It stores no IP address, user agent, referrer, contact information, employer or free text in this database. A key does not prove independent ownership. Our own and synthetic use is separated from verified external activity.
Application completions and aggregate counters expire after 90 UTC dates. The credential and first-use dates expire after 180 days without accepted use; daily maintenance runs without requiring new traffic. DELETE /v1/applications/current with your Bearer key revokes it and deletes linked events and cohorts. Anonymous aggregate counts remain. Application identifiers, token hashes and usage databases are excluded from evidence backups. No separate telemetry call is made by the daily job or local agent; an explicitly supplied key attributes their ordinary API requests. Enrollment, limits and deletion details.
Optional Research Desk measurement
The browser Research Desk runs without an account or measurement key. Its ordinary API query sends only your selected public dataset, entity filter, observation dates and protocol metadata. Do not enter confidential information in a filter. Anonymous response counters contain only UTC day, traffic class, outcome and count.
If you enable optional measurement, the browser stores a random installation key and sends it in the Authorization header on future research requests. The separate private store retains only a hash of that key, a random installation ID, dates, classification and hashed response identities. It does not retain filter text, IP addresses, user agents, referrers, names, employers, portfolios or advertising identifiers. The key represents an installation, not a verified person. There is no separate click, download or sharing event.
Browser enrollment stops after 30 days unless you enable it again. Unchecking the option disables future attribution. “Disable and delete my measurement” revokes the key and removes linked responses and cohorts. If deletion fails, the browser retains the disabled key so you can retry. Server responses and aggregate counts expire after 90 UTC dates; inactive credentials and cohorts after 180 days. Daily maintenance performs expiry even without new traffic. This store is excluded from public-evidence backups. Anonymous aggregate counts and ordinary infrastructure delivery logs remain separate.
Shared links contain only the selected dataset, filter, dates and fixed campaign labels. They contain no measurement key or saved observations. Share only filters you intend to make public. Downloads are prepared locally in the browser. Measurement definitions and limits.
Retention and contact
The evidence router does not persist topic request bodies. The funding service retains public research packets, public original-publisher responses and operational capture receipts in an append-only archive and encrypted recovery backups. These public evidence records are separate from optional usage counters; deleting a usage identifier does not delete public evidence. LIQUILENS PRIVATE LIMITED deletes operational logs it controls when the related security or diagnostic need closes, and no later than 30 days after collection, unless a longer period is required by law or to investigate active abuse. Provider-managed security and delivery logs follow each provider's configured lifecycle and privacy policy. For privacy questions, deletion requests, or a request concerning data controlled by LIQUILENS PRIVATE LIMITED, email mrinal@liquilens.in.