Automated processing
AI Disclosure
Thought Pins uses AI to help classify journal messages, distinguish conversation from saved memory, extract entities and relationships, summarize permitted sources, retrieve relevant context, and generate answers.
What AI does
Automated systems may route a message or user-initiated voice note as chat, journal, reminder, idea, expense, or source intent; transcribe audio; create structured memory from an entry; identify people, places, projects, events, things, and concepts; connect related records; summarize documents; and compose responses from retrieved context.
Human control
Users decide what to submit, save, import, export, and delete. AI-assisted routing can be corrected with undo, explicit save or chat instructions, content deletion, and account export or deletion workflows. When a message is ambiguous, the product may disclose whether it is replying as chat or saving a journal memory.
Voice notes and personal archive
A voice note is recorded only after the user chooses the microphone control and receives an in-app disclosure and operating-system permission prompt. The recording is transcribed and discarded by default. Personal voice archive is disabled in the current public store configuration. Private or self-hosted deployments may enable it as a separate choice that requires explicit acknowledgements before future recognizable recordings are retained in compressed, encrypted form.
Where enabled, retained recordings are limited to account-scoped personal voice functionality. No personal voice model is created in the current public release, and retained recordings are not used to train a shared model, another user's model, or a general-purpose foundation model. Users can stop future retention and permanently delete retained recordings independently of their journal transcripts.
Memory and source boundaries
Thought Pins is designed to keep personal journal memories distinct from articles, books, and documents a user has read. Generated answers may use both when relevant, but should preserve provenance so a source claim is not presented as something that happened to the user.
Limitations
AI responses and extracted memories can be incomplete, outdated, misclassified, or wrong. Retrieval may omit a relevant entry, and generated text may infer more than the underlying record supports. Users should inspect original entries and sources before relying on important details.
Provider-neutral runtime
The backend is provider-neutral. A deployment may use a configured AI provider for chat, routing, extraction, and summaries; an embedding provider for memory search; a local or disclosed hosted processor for voice transcription; article-fetch providers for user-submitted URLs; and crash-reporting providers for diagnostics. Provider processing is described in the Privacy Policy and store data-safety inventory. Production clients disclose the active processing categories before users submit personal content.
Safety and reporting
Users can report unsafe AI output, harmful or illegal content, copyright concerns, privacy concerns, or behavior that appears deceptive, harassing, exploitative, or dangerous through the in-app reporting control or support@thoughtpins.com. Thought Pins is not intended to generate malicious code, sexual exploitation content, deceptive election content, scams, harassment, self-harm instructions, or other prohibited content.
Privacy
AI and embedding providers may receive the content necessary to perform a requested classification, extraction, search, summary, or response. The Privacy Policy describes provider categories, consent, retention, user rights, and account deletion.
No professional advice
Thought Pins is a journaling and memory product. It does not provide medical, legal, financial, emergency, or other professional advice. Contact qualified professionals for decisions requiring professional judgment.