# Blueprint: Sponsor inbox

URL: https://insidethemachine.ai/build/sponsor-inbox
Part: B-04. Family: Creator ops. Status: Blueprint, not run yet.
Blueprint. Not run yet. Local Jev, the sorting part, runs today (tested on a small model on a cloud CPU).

Sort every sponsor email on your own machine. Real offers rise to the top, scams and mass pitches drop out. Nothing leaves your computer.

- Input: Sponsor emails saved as text files in one folder.
- Output: A sorted list: real offers first, each with a fit score and a reply flag. Draft replies for the clear ones.
- Runs from: 8 GB of memory.
- Tools: ollama (https://insidethemachine.ai/setups/ollama.md)

## Claude's job

- Writes the replies to the best offers only, from a short summary you approve first.
- Never sees the other emails.

What Claude never sees: Every email is sorted on your computer by Local Jev. Claude sees only the summaries you choose to send, after a preview.

## Local parts per memory size

Data: models.json 2026-10-15, speech.json 2026-10-15, images.json 2026-10-15.

| memory GB | Decider (Local Jev) | Drafter |
|---|---|---|
| 8 | Qwen3.5 4B (needs 5.8 GB, runs, but tight) | Qwen3.5 4B (needs 5.8 GB, runs, but tight) |
| 16 | Qwen3.5 0.8B (needs 2.6 GB, runs well) | Gemma 4 12B (needs 11.2 GB, runs, but tight) |
| 24 | Qwen3.5 4B (needs 5.8 GB, runs well) | Gemma 4 12B (needs 11.2 GB, runs well) |
| 32 | Qwen3.5 4B (needs 5.8 GB, runs well) | Qwen3.8 27B (needs 22.7 GB, runs, but tight) |
| 48 | Gemma 4 12B (needs 11.2 GB, runs well) | Qwen3.8 27B (needs 22.7 GB, runs well) |
| 64 | Gemma 4 12B (needs 11.2 GB, runs well) | Qwen3.8 27B (needs 22.7 GB, runs well) |
| 96 | Qwen3.8 27B (needs 22.7 GB, runs well) | Qwen3.8 27B (needs 22.7 GB, runs well) |
| 128 | Qwen3.8 27B (needs 22.7 GB, runs well) | Qwen3.8 27B (needs 22.7 GB, runs well) |
| 192 | Qwen3.8 27B (needs 22.7 GB, runs well) | Qwen3.8-Flash-Next (125B-A6B) (needs 140 GB, runs, but tight) |
| 256 | Qwen3.8 27B (needs 22.7 GB, runs well) | Qwen3.8-Flash-Next (125B-A6B) (needs 140 GB, runs well) |
| 512 | Qwen3.8 27B (needs 22.7 GB, runs well) | Qwen3.8-Flash-Next (125B-A6B) (needs 140 GB, runs well) |

How each part is picked:
- Decider (Local Jev): Answers typed questions about each item: pick one, give a score, or yes/no. It never writes. Rule: The smartest chat model with an Ollama download that needs no more than a third of your usable memory. Smaller models answer sooner, and a decider only picks a label. If no model is that small, the drafter's model decides too. (models.json)
- Drafter: Writes the first draft of many small things: summaries, title options, reply drafts. Rule: The smartest chat model that fits your usable memory and has an Ollama download (the same model the /do recipes install). (models.json)
- Fit: We count 75% of your memory as usable for AI. The rest is for the system and your apps. Runs well: the model fits in usable memory with at least 10% to spare. Runs, but tight: it fits, but less than 10% of usable memory is left. Close other apps first. Needs N GB+: it does not fit. N is the smallest memory size it fits in. Size unknown: we do not know how much memory it needs.

## Steps

1. **Save the emails to a folder** (You; stage id `export`): One text file per email. They stay on your computer.
2. **Strip the noise** (A script; stage id `clean`): A script removes signatures, quoted replies and tracking links.
3. **Local Jev sorts them** (Your machine, Decider (Local Jev); stage id `triage`): Pick one: real sponsor, agency mass pitch, scam, fan mail, other. Score: how well does it fit your channel? Yes/no: does it need a reply? Unsure answers come to you.
4. **Draft the easy replies** (Your machine, Drafter; stage id `drafts`): Polite no-thanks drafts for clear mismatches. Nothing is sent.
5. **Claude writes the important replies** (Claude; stage id `replies`): For the top offers only, from a short summary. You see exactly what goes to Claude first.
6. **You send** (You; stage id `yours`): Read, edit and send yourself. Never let a verdict alone send, delete or click anything.
7. **Share your run (optional)** (You; stage id `share`): At the end, the script shows exactly what it would upload: your chip, memory, the models, how long each stage took and the token counts. No prompts, no outputs, no file names. Say yes and it goes to our review queue. Nothing is public until we approve it.

## Make it yours

- Emails can hide instructions aimed at AI. A verdict never sends, deletes or opens a link on its own: you do.
- Check every offer and payment detail yourself.

## Receipts

- Claude tokens, this recipe: not run yet
- Claude tokens, all in Claude: not run yet
- Local minutes: not run yet
- Claude dollars (estimate): not run yet

No approved runs yet. Blueprints now, numbers later. We run every recipe on a 128 GB Mac Studio from 10 Nov 2026, then publish the receipts.

## Share your run

Opt-in. The script shows exactly what it would upload and asks first. It sends: Your chip, memory, graphics card and operating system; The recipe and the models you used; How long each stage took; Token counts: Claude in and out, local in and out; How many items came out; Tool versions; A fingerprint (SHA-256) of the log and up to 15 cleaned log lines; A handle you pick (not your name or email). Never: Your prompts or the model's answers; Your files, file names or folder paths; Emails, links, keys or IP addresses (the server rejects any upload that has one); Your IP address in plain form: we keep a scrambled copy for rate limits only. Nothing is public until a person approves it. Agents can submit with the `submit_run` tool of https://insidethemachine.ai/mcp.

Privacy: https://insidethemachine.ai/privacy.md