The short version: Your documents never leave your infrastructure. We do not store, process, or transmit your client data to any external server. Zero Data Retention is not a policy — it is the architecture.
Fletchr is built on a Zero Data Retention (ZDR) architecture. This means that documents uploaded to your Fletchr instance are processed and stored exclusively within your private infrastructure — not on Fletchr servers, not on third-party cloud servers, and not on any external system.
Your vector embeddings, document chunks, and query history remain within your deployment environment. When you query your documents, the request is processed locally and responses are generated without your document content leaving your infrastructure.
Fletchr collects only the minimum information necessary to operate the service:
— Account information: your email address and authentication credentials, stored securely in Supabase with industry-standard encryption.
— Usage metadata: aggregate query counts, document counts, and storage usage for billing and service monitoring purposes. This metadata does not include document content.
— Communication data: emails you send to mark@getfletchr.com for support or sales purposes.
— The contents of any document you upload to your Fletchr instance.
— The text of any query you run against your documents.
— Any client information, case information, or privileged material contained in your documents.
— Vector embeddings or semantic representations of your documents.
Fletchr uses the Anthropic Claude API with Zero Data Retention enabled. This means that prompts and responses processed through the Claude API are not stored or used to train Anthropic's models. Each session is processed and immediately discarded.
For clients requiring additional assurance, Fletchr offers fully local AI processing through AWS Bedrock as a fallback option, eliminating any external API dependency.
Chronicle is a separate product from Fletchr. Chronicle indexes publicly available government documents — federal regulations, court opinions, congressional records, and declassified materials. No private user data is indexed in Chronicle.
Chronicle accounts are governed by separate terms. Query history on Chronicle is stored to provide the search history feature and may be used in aggregate to improve search quality.
Your Fletchr instance runs on isolated infrastructure with the following security controls:
— End-to-end TLS encryption for all data in transit.
— Row-level security policies enforced at the database level via Supabase.
— MinIO object storage with access controls scoped to your firm ID.
— ChromaDB vector database bound to localhost only — not accessible from external networks.
You may request deletion of your account and all associated metadata at any time by contacting mark@getfletchr.com. Because your document data is stored within your own infrastructure, deletion of your deployment environment removes all document data without any action required from Fletchr.
Privacy questions, data requests, or security concerns:
Fletchr LLC · North Carolina