Nanosaur2-670M
well9472 · Kit
Makes illustrations from text prompts with a small 670M-parameter diffusion model, run through ComfyUI. Built as a low-budget research model for people who wan…
A small n8n tool workflow that another workflow or an AI agent calls with a request such as unread emails from yesterday. OpenAI works out the date, Gmail fetches up to ten unread inbox emails since then, and OpenAI summarizes each one.
Summarize unread Gmail by date is an n8n workflow in the Productivity category, listed by Mindket. It is free. This is version 1, last updated on . Download it on this page or copy its JSON, then import it into an n8n canvas. Nothing runs until you activate it in n8n.
Call from another workflow
It starts when another workflow calls it, passing:
A short summary of each unread email since that day, handed back to the caller
What you need before you start.
Language model
Works out the date from the request, then summarizes each email with its sender.
NeedsOpenAI API key
Fetches up to ten unread inbox emails received after the date, in full.
NeedsGoogle sign-in for Gmail
Checked by Mindket on 13 Nov 2026
This n8n workflow is a tool for other workflows to call, and it only reads. Whatever calls it passes one input named query, a request that names a day. OpenAI (gpt-4o) reads it with today’s date and answers with one date. The Gmail node fetches up to ten emails that are in the inbox and unread, received after that date. A second OpenAI step summarizes each email with its sender, Aggregate gathers the summaries into one list, and the last node sets response. Six nodes in all.
People who want to ask an assistant what they missed, and builders of n8n agents that need an inbox tool for it. Skip it if you need every unread email or a folder other than the inbox. It stops at ten messages and filters on the inbox label.
Inside a larger workflow, as the target of an Execute Workflow node or as a tool an AI agent can call. Any n8n works, as long as a Gmail account can be authorized in it.
Get it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Its author puts setup at about 10 minutes.
workflow.json.OpenAI and OpenAI1.Gmail.When Executed by Another Workflow and check the field’s name. The file has query followed by a stray line break, so delete the break and let it read exactly query.Gmail: the date should be there (see Good to know).The caller passes one field and gets one back:
{
"query": "unread emails from yesterday"
}{
"response": "<the summaries, one for each email>"
}Ask in plain words and name a day. For today or yesterday the date step returns midnight UTC of that day, and with no day named it uses today. The Gmail filter means received after that moment, so a request for yesterday returns every unread inbox email from the start of yesterday until now, today’s included. Each summary starts with the sender’s name and then gives a short account of the plain-text body.
A caller passes the request from the fence above. The date step answers with yesterday at 00:00:00Z, Gmail returns up to ten unread messages since then, and the caller gets one short summary for each, headed by who sent it.
If you are new. A tool workflow runs only when another workflow calls it. Unread and inbox are Gmail labels, and the workflow filters on both, so a message you have already opened is not included. Nothing is marked read or changed.
For experts. Gmail runs getAll with a limit of 10, label IDs INBOX and UNREAD, and simplify turned off, which returns the full message that the summary step needs for the sender and the body. Both OpenAI nodes use gpt-4o. The date step is told to answer with ISO 8601 and nothing else. The filter reads $json.date and falls back to the current time when that is empty.
Check the date reaches Gmail. The risks after that are the cap and the privacy of your inbox.
message.content, while the filter reads $json.date. If the date is not at the top level of Gmail’s input, the filter falls back to the moment of the run and finds nothing. Point receivedAfter at {{ $json.message.content.date }}An n8n workflow, and a tool for other workflows to call. Give it a request that names a day and it returns a short
summary of each unread inbox email since then.
workflow.json: the workflow as n8n exports it, six nodes. It holds no credential.README.md: these notes.workflow.json.OpenAI and OpenAI1.Gmail.When Executed by Another Workflow. The file names its one field query followed by a stray line break.query.{ "query": "unread emails from yesterday" } pinned on the trigger. Open the input of Gmail| Direction | Field | What it holds |
|---|---|---|
| In | query | a request that names a day, such as unread emails from yesterday |
| Out | response | the summaries, gathered into one value, one for each email |
OpenAI gets the request and the current date and answers with one date in ISO 8601, as {"date": ...}. Today andGmail fetches up to 10 messages with the labels INBOX and UNREAD received after that date, with simplify off so theOpenAI1 writes a short summary of each one, headed by the sender's name.Aggregate gathers the summaries and Edit Fields puts them in response.Gmail.OpenAI.OpenAI1.message.content, while the filter in Gmail reads$json.date. If the date is not at the top level of the Gmail node's input, the filter falls back to the moment ofReceived After at {{ $json.message.content.date }}.README.md 3 KB workflow.json 5 KB
well9472 · Kit
Makes illustrations from text prompts with a small 670M-parameter diffusion model, run through ComfyUI. Built as a low-budget research model for people who wan…
Waqf-AI · Kit
Reads handwritten Arabic text from images and returns it as text. A fine-tuned PaddleOCR-VL model for people digitizing Arabic handwriting.
Wan-AI · Kit
Makes video clips from a text prompt or an image plus text, at up to 720P and 24 frames per second, in the Diffusers format for a single consumer GPU.
Wan-AI · Kit
Generates short 480P videos from a text prompt with the small 1.3B Wan2.1 model, packaged for the diffusers library. Runs on a consumer graphics card with abou…
serendipitylabs · Kit
Turns spoken Malayalam into text on a CPU or a CUDA GPU, as an early research release from Serendipity Labs. For people who want draft transcripts of short, cl…
OrdenWills · Kit
Turns plain home commands such as turn off the TV into structured tool calls, checking device states first. A 350M model for smart home builders and tinkerers.