Note: each example names which recent release or tracker it leverages so you can try the workflow using the appropriate provider or an equivalent service.
High‑quality oral history transcription using MAI‑Transcribe-1 — record interviews on your phone, upload to a transcription endpoint or Microsoft OneDrive, run MAI‑Transcribe‑based transcribe, then export time‑stamped text for indexing and source citations.aiflashreport
Create narrated family history videos with MAI‑Voice-1 — feed transcribed scripts to MAI‑Voice‑1 to generate consistent narration styles for YouTube or family DVDs.aiflashreport
Handwritten note transcription + clean‑up (Perplexity + local LLM) — screenshot handwritten research notes, run OCR, then use a small local open model (MiniCPM or Granite small) to normalize names/dates and propose structured facts.aiflashreport
Batch census‑to‑timeline extraction (Perplexity research pages + local model) — feed census images and enumerations to Perplexity’s research/citation flows, extract person timelines, and save to your research log.marketingaiinstituteyoutube
Quick source finder with Perplexity Pages — give a surname/place pair to Perplexity Pages to generate a shareable reference page listing where to look next (repositories, search terms, likely record types).youtubemarketingaiinstitute
Local privacy workflow with Granite or MiniCPM — run sensitive family data through a local Granite/MiniCPM instance to produce summaries without sending data to a cloud provider.aiflashreport
Rapid FOIA / archive request drafting using Claude/OpenAI agent concepts — use a “coordinator” style prompt (one agent drafts, another validates dates/places) to produce clean archive request letters and citation lists (mirrorable in Claude or ChatGPT when agent features roll out).youtube
Source evaluation checklist generator (Perplexity + local LLM) — paste a transcription or record image and ask for a Greenwood‑style source evaluation checklist with confidence flags using Perplexity’s citation context.marketingaiinstituteyoutube
Newspaper search expansions using Perplexity research + MAI transcription — transcribe long obituary audio or old phonograph recordings, then have Perplexity suggest newspaper search terms and date ranges.youtubeaiflashreport
DNA segment narrative with genomics AI output — where you have research‑grade coalescence estimates, use the new DNA coalescence tool (research pipeline) to create time‑to‑common‑ancestor narratives for known matching segments, then write an interpretive paragraph for client reports.earth+1
Bulk vital‑record citation creator (local Granite + Zotero) — feed extracted facts from records into a local Granite model to format Zotero‑ready citations and research notes.aiflashreport
Handwriting verification runlist — use a small open model to compare multiple handwritten samples (signature, margin notes) and produce a likelihood score for same‑writer matches as a triage step before expert paleography review.aiflashreport
Multi‑record reconciliation agent (agent/coordinator pattern) — implement a two‑stage workflow where one model normalizes names/dates and a second model cross‑validates conflicts and suggests next searches (pattern coming from agent roadmap discussions).youtube
Quick repository query templates — generate tailored FOIA/archive query templates (subject, date range, fees, contact) using Perplexity or a local open model for the repository you plan to contact.youtubeaiflashreport
Family narrative drafts from timelines (Perplexity + MAI voice) — turn a structured timeline (from Perplexity/LLM extraction) into a readable narrative and synthesize it into an audio narration via MAI‑Voice‑1 for client listening files.youtubeaiflashreport
Transcribe and synchronize audio with records — transcribe oral histories with MAI‑Transcribe‑1, then annotate transcript with record citations and timestamps so clips align with referenced documents.aiflashreport
Newspaper clipping summarizer for client packets — feed multiple newspaper clippings into Perplexity to create a one‑page summary with suggested research leads and citation links.youtube
Brick‑wall brainstorming session (Perplexity + local LLM) — ask Perplexity for focused research hypotheses, then run a local model to convert top hypotheses into prioritized search tasks for your next repository visit.youtubeaiflashreport
Create workshop handouts and prompts — use Perplexity Pages to generate a shareable one‑page handout (search checklist + prompts to use in ChatGPT/Claude) for your next genealogy class.marketingaiinstituteyoutube
Sensitive client data redaction workflow — process documents locally with Granite or MiniCPM to redact personal data before sharing research drafts externally.aiflashreport
Batch name‑variant generator tuned to locality — ask a local/mini model to produce historical spelling variants and immigration phonetic variants for a surname in a given county/year range to expand search queries.aiflashreport
Quick transcription quality audit — run MAI‑Transcribe output and a small open model to compare transcriptions and flag low‑confidence segments needing human review.aiflashreport
Citation‑first note production — produce notes where each fact is followed immediately by a formatted citation (Perplexity’s citation flows), ready to paste into Zotero notes.marketingaiinstituteyoutube
Audio interview voice matching for family members — create custom voices from short audio samples with MAI‑Voice‑1 to produce consistent narrated readings of letters or diaries.aiflashreport
One illustration/example: To produce a narrated family biography for a client this afternoon — (1) record or collect audio clips and scanned documents; (2) transcribe interviews with MAI‑Transcribe‑1; (3) extract timeline and facts with Perplexity Pages; (4) clean and format citations with a local Granite/MiniCPM instance; (5) generate narration with MAI‑Voice‑1 and assemble a short MP4. The same pipeline can be used to produce short narrated case studies for teaching.
No comments:
Post a Comment