Truly private AI.
Late one?
The longer version: ChatVLM is a complete AI workspace: chat, documents, research and agents, with a private pool of models behind it. It runs where your data already lives, so nothing sensitive has to leave the building to get answered.
Board update · November
Q3 at a glance
+18%
Revenue, quarter on quarter
1 / 8
Eight slides, charted from your numbers. It’s versioned in the side panel, so ask for changes and the earlier version stays.
| A | B | C | D | E | |
|---|---|---|---|---|---|
| 1 | Q4 budget | ||||
| 2 | Category | October | November | December | Q4 total |
| 3 | Staff | 48,200 | 48,200 | 51,600 | 148,000 |
| 4 | Ingredients | 21,400 | 23,900 | 27,800 | 73,100 |
| 5 | Rent | 9,500 | 9,500 | 9,500 | 28,500 |
| 6 | Equipment | 2,100 | 14,800 | 1,200 | 18,100 |
| 7 | Marketing | 3,000 | 4,500 | 6,200 | 13,700 |
| 8 | Utilities | 2,900 | 3,100 | 3,600 | 9,600 |
| 9 | Packaging | 1,700 | 1,900 | 2,400 | 6,000 |
| 10 | Total | 88,800 | 105,900 | 102,300 | 297,000 |
| 11 | |||||
| 12 | |||||
| 13 | |||||
Memo
Memo to the board
- To
- The board
- From
- Finance
- Re
- Q4 budget
Q4 spending comes to $297,000, three per cent under plan. Staff stays the largest line at half the total, and the new oven lifts equipment in November only.
What we ask the board to approve
- The Q4 budget as attached, with December’s seasonal hires.
- A one-off equipment line for the oven, paid in November.
- Moving the marketing increase into December.
Ingredients rise with the holiday orders and fall back in January. Rent, utilities and packaging hold close to plan.
Three more, made in the conversation: a sheet you can sort and chart, a one-page memo in Word, and a cover for the deck.
A four-hour response on priority incidents, around the clock, and a service credit for every hour past it.
Halden MSA.pdf page 14 “Supplier shall respond to Priority 1 incidents within four (4) hours, twenty-four hours a day…”
Yes. They speak the standard APIs these tools already use, so the only thing that changes is the base URL.
BASE_URL=https://ai.your-company.internal/v1
On a Sovereign Box the meter is off. The tenth agent costs what the first one did.
Ours, to start. Every plan, free included, carries zero data retention in the contract, and nothing you send trains a model.
A private cloud. Sovereign Presets run open-weight models on infrastructure operated for you alone, with zero third-party subprocessors and no rack to install.
Your building. A Sovereign Box serves your people from your own rack, with unlimited seats and usage, and keeps answering when the big providers go down.
It turns on where your data is allowed to go. If it can’t leave the building, a Sovereign Box. If it can leave but must stay in your control, Sovereign Presets in a private cloud. If neither, start in ours today.
Here’s the receipt. Tell me where this ran, and I’ll reprint it.
- Your messages
- 7
- Creations
- 4
- Files read
- 2
- Council drafts
- 3
- Kept in your workspace
- Until you delete it
- Sent to the model
- Only what each answer needed
- Kept by the model provider
- Nothing, dropped when each request completed
- Browsed by our staff
- No
- Used to train a model
- Nothing
Closing your account deletes all of it immediately, with no grace period and no archived copy. We keep billing records and usage analytics, and staff read content only for support you ask for, abuse, or the law. Privacy Policy, sections 2, 5 and 8.
Nothing retained. Nothing trained on. It’s in the contract on every plan, the free one included.
Free to start. Every plan, free included, comes with zero data retention in the contract and no training on your data.
Yes. One of us will read it and get back to you personally, usually within one business day. I’ve drafted it from this conversation, so change anything you like.
ChatVLM can make mistakes. Check important info.