πŸ“²
VisionIT

Talk to Vision

Record a thought. Build the company's memory.

Use your VisionIT Microsoft 365 account.

Developer mode β€” no password. Configure Microsoft sign-in in Settings for production.

VisionIT
Talk to Vision

Talk to VisionIT

Tap the mic and talk β€” it transcribes and saves to your Drafts to review later.

Tap to record
0:00 / 2:00

Recording stops automatically at 2:00. For longer audio, upload a file instead.

Talk to VisionIT

Ready β€” tap & talk
🌐

✍️ It writes as you speak β€” or just type Β· edit before saving

πŸ”” Remind me (24h Β· only you)
πŸ“š Share with the team
πŸ”’ Personal
πŸ”· Azure DevOps

Drafts β€” review before publishing

Everything you capture β€” audio, transcripts, and photo/video notes β€” lands here as a private draft. Review it, then either Publish it as a note (pick a category) or βž• Make Azure task if it's actually a task. Nothing is shared until you publish. Add new captures from the Upload screen.

No drafts to review.

βœ… Azure tasks

πŸ”— Open my tasks in Azure β†—

Describe a task (or paste what you said) β€” e.g. "task to fix the EIR accrual for Deloitte NBG, assign to Maria". It's parsed into a draft; pick Client β†’ Project β†’ Epic, fill the required fields, then Create in Azure. Nothing is created until it's complete and you approve. Once created, a task moves to Azure β€” open the board above to see all your created tasks.

My Journal

Private to you β€” only πŸ“˜ Knowledge is shared into Ask VIT.

To-Do Idea Knowledge Azure

Ask VIT

Ask a question β€” the AI VIT Consultant reads the company knowledge base (notes, trainings & uploads) and answers. Ask follow-ups β€” the conversation keeps its context. Open Sources under any answer to see exactly what it used.

Mode:

πŸ“€ Upload Knowledge

Add files you already have. Audio, photos & videos are transcribed/read by AI and land in Drafts for review; documents (transcripts, PDFs, decks) are added as searchable knowledge. To record live, use the Record tab.

🎧 Upload audio

Drop in a voice memo β€” transcribed on-device by local Whisper (offline) and saved to Drafts.

Pick the language you're speaking β€” Auto-detect can misfire on short clips.


πŸ“Έ Photos & video

Upload photos β€” or a short video of pages being turned β€” of handwritten or printed notes. AI reads the text; review each in Drafts.

For an older notebook, pick the year β€” pages usually print only the weekday & day.

πŸ“ Paste or upload a transcript

A meeting/call transcript or any text β€” saved to Drafts for review.


πŸ“Š Presentations & PDFs

PowerPoint decks (.pptx) or PDFs. Each slide/page's text becomes searchable knowledge β€” shown under Uploads in Ask VIT.

πŸ“¦ SharePoint Trainings Status

Tracks the VisionIT Team Trainings & Useful Material SharePoint library against what has been converted to Markdown and uploaded here β€” so you can be sure nothing was missed.

Loading…

Manage knowledge

Edit or remove πŸŽ“ imported trainings and πŸ“Š presentation / PDF uploads. Notes are never shown here β€” nobody can read another person's notes; journal & personal entries belong to their owners.

Loading…

Settings

Your personal preferences on this device.

🎨 Appearance

Theme for this device.

β€œAuto” follows your operating-system light/dark setting.


πŸš€ Startup screen

Which screen opens when you sign in on this computer.


ℹ️ Welcome tour

Replay the short intro that new users see on first login.


🎀 Hands-free capture (Siri Shortcut)

Capture a note by voice from anywhere β€” including in the car with CarPlay ("Hey Siri, Note to Vision"). The dictated text lands in your Drafts for review.

⚠ The token is personal β€” anyone with it can add drafts to your journal. Don't share it.

  1. On your iPhone open Shortcuts β†’ + new shortcut β†’ name it Note to Vision.
  2. Add action Dictate Text (set language as you prefer).
  3. Add action Get Contents of URL β†’ set the URL above Β· Method POST Β· Request Body JSON with two fields: token = your token Β· text = the Dictated Text variable.
  4. Say "Hey Siri, Note to Vision" β€” works hands-free, incl. while CarPlay is connected.

Admin Settings

App-wide configuration and administration. Admins only.

πŸ›‘οΈ Admins

Admins can open Manage & Admin Settings and see everyone's usage. You can't remove yourself.


πŸ‘₯ Who's using the app

People with an active session. 🟒 = active in the last 5 minutes.


βš™οΈ Technical settings

Connections & keys β€” set up once, centrally. Most people never need to open these.

Single sign-on β€” Microsoft Entra ID

Let staff sign in with their VisionIT Microsoft 365 account (no passwords, inherits company MFA). Status: …. Setup steps: deploy/AUTH_SSO_SETUP.md.

Register a web app in Entra admin center β†’ App registrations (single tenant), add the redirect URI above, create a client secret, then paste the three values here. Leave the secret blank to keep the saved one. Once saved, the passwordless dev login is disabled. Env vars: ENTRA_TENANT_ID/ENTRA_CLIENT_ID/ENTRA_CLIENT_SECRET/ENTRA_REDIRECT_URI.


Microsoft Graph β€” SharePoint trainings

Lets the app read the SharePoint trainings library: the πŸ”„ Re-scan button and the per-file βž• Add buttons. Status: ….

Needs an Entra app registration with the application permission Sites.Read.All (or Files.Read.All) and admin consent β€” that consent step requires a Microsoft 365 admin. You can reuse the same app registration as SSO above (just add this permission + a secret). Leave the secret blank to keep the saved one. Env vars: GRAPH_TENANT_ID/GRAPH_CLIENT_ID/GRAPH_CLIENT_SECRET.


AI search

AI semantic search & Q&A use Claude. Status: ….

Stored locally in data/config.json on the server machine. For production, prefer the ANTHROPIC_API_KEY environment variable.


Azure DevOps Boards

Create work items by voice from the Record screen. Status: ….

Create a PAT in Azure DevOps β†’ User settings β†’ Personal access tokens, scope Work Items (Read & Write). Stored locally in data/config.json; for production prefer the AZURE_DEVOPS_PAT environment variable.


Voice transcription engine

Used for audio in Drafts. Status: ….

Local Whisper runs on this machine β€” no key, no cost, audio never leaves the company. Bigger models = better accuracy but slower on a CPU. Changing the model downloads it on first use. Azure is the cloud alternative below.


Azure AI Speech (cloud alternative)

Only needed if Engine = Azure above. Status: ….

Create a Speech resource in the Azure Portal (free tier: 5 hours/month). Supports Greek & English. Env vars: AZURE_SPEECH_KEY, AZURE_SPEECH_REGION.

Talk to Vision Β· prototype Β· VisionIT