Huashu Design
alchaincyf · Skill
Turns a one-line request in your AI agent into a finished design: a launch animation, a clickable app prototype, an editable slide deck, an infographic or a se…
An n8n workflow for Cal.com bookings. OpenAI models read the booking notes and research the person and their company on the web. Gmail then sends you a short brief with an MP3 of it read aloud.
Emailed meeting brief with audio for each Cal.com booking is an n8n workflow in the Business 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.
New Cal.com booking
It starts when someone books a meeting with you on Cal.com, with:
A meeting brief in your inbox for each booking, as bullet points with an MP3 to play
What you need before you start.
Scheduling
Sends each new booking to the workflow through a webhook that n8n registers.
NeedsCal.com API key
Language and speech models
Runs gpt-5.6-sol, gpt-5.6-terra and gpt-5-search-api for the analysis, the research and the writing, and tts-1 for the audio.
NeedsOpenAI API key
Sends the brief and its MP3 to the address you set.
NeedsA Google account to sign in with
Checked by Mindket on 13 Nov 2026
This n8n workflow prepares you for meetings that people book with you on Cal.com. Each new booking starts it. One model reads the booking notes and works out what the person probably wants, while gpt-5-search-api looks up the person on the web and, in a second call, their company. A fourth call joins the three answers into one short brief, written to be heard. It reaches you by Gmail as an HTML email, with an MP3 of the brief attached. Eighteen nodes do the work, under a sticky note that titles the canvas.
Consultants, agencies and founders who take calls from strangers through a Cal.com link and want to know who booked before they join. It pays off on a calendar too full to prepare by hand. Skip it if your bookers are colleagues. The prompts assume a client call.
| Industry | Who uses it | How they use it | Example |
|---|---|---|---|
| AI and automation agencies | Founders who take sales calls | Keep the prompts, which are written for an AI agency | A two-person agency plays the audio before each first call with a new lead. |
| Consulting | Independent consultants | Ask for the company on the booking form, and read the questions it suggests | A consultant listens to the brief on the way to a client’s office. |
| Recruiting | Recruiters | Rewrite the person prompt to cover a career and public work | A recruiter reads a short profile of each candidate before the screening call. |
| Coaching | Business coaches | Use the notes field to ask what the client wants to change | A coach reads the likely goals and three questions before a first session. |
You need some n8n and three accounts: Cal.com, OpenAI and Gmail. One booking question, three credentials and your email address make up the setup. Its author puts it at about 12 minutes.
Any n8n that Cal.com can reach with a webhook will run it, so an n8n on your own laptop needs a public address or a tunnel before the trigger hears about any booking. The briefs arrive wherever you read Gmail.
YOUR_EMAIL so whoever takes the call gets the briefGet it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Have these ready before you import it.
| Need | Detail |
|---|---|
| n8n | On n8n Cloud, or on a server Cal.com can reach |
| Cal.com | An API key, and a booking question with the identifier Company-Name |
| OpenAI | An API key for gpt-5.6-sol, gpt-5.6-terra, gpt-5-search-api and tts-1 |
| Gmail | An account to send from, for a Gmail OAuth2 credential |
workflow.json.Cal.com Trigger.Meeting Goal Analysis, Meeting Prep Report Generator, Meeting Brief Email Formatter, Generate Audio Briefing, OpenAI Chat Model and OpenAI Chat Model1.Gmail and replace YOUR_EMAIL with the address that should get the briefs.Company-Name.There is nothing to press. Once the workflow is published, each new booking starts a run on its own, and the email follows as soon as the last model call and the audio are done. Read the brief or play the MP3. The goal analysis only has the notes to go on, so ask bookers what they want to talk about.
| To change | Where |
|---|---|
| Who you are | The first line of Meeting Goal Analysis, which describes an AI agency that brings AI into its clients’ systems |
| What the research covers | The prompt text of AI Agent for the person, and of AI Agent3 for the company |
| Length and shape of the brief | The prompt of Meeting Prep Report Generator, which asks for under 1,500 characters |
| The email | The prompt of Meeting Brief Email Formatter, and the subject in Gmail |
| The voice | Generate Audio Briefing, on the node’s defaults: tts-1, the voice alloy and MP3 |
Tomás Reyes books a call. He answers Harbor Dental Group to the company question, and his note says he wants fewer missed appointments.
| Part | What it holds |
|---|---|
| Subject | The words Meeting Prep and his name |
| Body | Bullet points in HTML: who he is, what Harbor Dental Group does and any recent news, what he probably wants, and a few questions to ask |
| Attachment | An MP3 of the full brief |
| Last line | n8n’s note that the email was sent automatically |
The search model’s only source is public pages. Someone with little public presence gets a thin profile, so check names and titles before you repeat them on the call.
If you are new. n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. This one starts from a Cal.com event, not a button. It splits into three branches for the notes, the person and the company, then joins them before the writing starts. Start with that test booking. Then open the run under Executions and click through the nodes to read each answer.
For experts. The two research steps are AI Agent nodes with no tools. Each sends one prompt to gpt-5-search-api, a Chat Completions model that searches the web by itself.
| Node | Model or service | What to know |
|---|---|---|
Meeting Goal Analysis | gpt-5.6-sol | Reads only the booking notes. The file named o3-mini, now retired |
AI Agent, AI Agent3 | gpt-5-search-api | The search happens inside the model, so n8n shows no search step |
Meeting Prep Report Generator | gpt-5.6-sol | The same swap from o3-mini. Its prompt asks for text that reads well aloud |
Meeting Brief Email Formatter | gpt-5.6-terra | The file named o4-mini. Simplify Output is off, so Gmail reads choices[0].message.content |
Generate Audio Briefing | OpenAI text-to-speech | Writes the MP3 to the binary field data, which Gmail attaches |
Assemble Email Package matches its three inputs by position, so it expects one booking per run, which is what the trigger sends.
Every new booking runs it, and each run sends the attendee’s details to OpenAI.
Extract Booking Details reads .value from the company answer and from the notes, so a booking without them can stop the run. Make the company question requiredo3-mini and o4-mini on 23 October 2026, and Mindket swapped in the replacements its deprecations page namestts-1 for removal on 6 January 2027, with gpt-realtime-2.1-mini as its replacement. Until then the audio step runs as it isGmail and had lost the backslashes in both research promptsAn n8n workflow for people who take meetings through Cal.com. When someone books, OpenAI models read the booking notes,
search the web for the person and their company, and write a short brief for you. The brief reaches your inbox through
Gmail as an HTML email with a spoken MP3 version attached.
workflow.json: the workflow as n8n exports it, 19 nodes, one of them a sticky note that titles the canvas. It holds noREADME.md: these notes.Gmail had no recipient, so the send step would have failed. Its To field now reads YOUR_EMAIL.o3-mini and o4-mini asMeeting Goal Analysis and Meeting Prep Report Generator nowgpt-5.6-sol, and Meeting Brief Email Formatter uses gpt-5.6-terra, the replacements that page names.AI Agent and AI Agent3 had lost their backslashes, so their \n line breaks had turned into a straygpt-5.6-sol, gpt-5.6-terra, gpt-5-search-api and tts-1.Company-Name. The workflow reads the company fromworkflow.json.Cal.com Trigger.Meeting Goal Analysis, Meeting Prep Report Generator,Meeting Brief Email Formatter, Generate Audio Briefing, OpenAI Chat Model and OpenAI Chat Model1.Gmail and replace YOUR_EMAIL with the address that should get theCompany-Name. TheCal.com Trigger fires on BOOKING_CREATED. Extract Booking Details keeps the first attendee's name and email,Company-Name answer and the notes.Meeting Goal Analysis reads the notes and suggests the booker's goals and questions toAI Agent asks gpt-5-search-api about the person, and AI Agent3 asks it about the company.Combine Research Results and Consolidate Research Data gather the three answers into one item.Meeting Prep Report Generator writes one brief meant to be heard, under 1,500 characters by its prompt, with a fewGenerate Audio Briefing reads the brief aloud with OpenAI's text-to-speech, and Prepare Audio Attachment passes theMeeting Brief Email Formatter turns the same brief into bullet points in HTML.Assemble Email Package lines up the booking, the HTML and the audio, and Gmail sends them to YOUR_EMAIL. TheMeeting Prep - followed by the attendee's name.Meeting Goal Analysis and Meeting Prep Report Generator. The first describesAI Agent (the person) and AI Agent3 (the company).Meeting Brief Email Formatter, and the subject in Gmail.Generate Audio Briefing uses the node's defaults, tts-1 and the voice alloy, as MP3. Pick anotherPrepare Audio Attachment if you want text only.Company-Name answer, or no notes, can stop the run at Extract Booking Details. Both expressions.value, so make the company question required.tts-1 for removal on 6 January 2027 and names gpt-realtime-2.1-mini as itsgpt-5.6-sol calls costREADME.md 5 KB workflow.json 16 KB
alchaincyf · Skill
Turns a one-line request in your AI agent into a finished design: a launch animation, a clickable app prototype, an editable slide deck, an infographic or a se…
op7418 · Skill
Skill that makes single-file HTML slide decks, slide illustrations and platform covers in two visual styles, with a presenter mode for rehearsal. For speakers…
alirezarezvani · Skill
Adds 411 ready-made skills for Claude Code and other coding agents, grouped by role from engineering and product to marketing, finance and compliance. Installs…
perplexity-ai · Kit
Picks the best option for a decision from text and images, such as which team should handle a request, with a probability for each choice, for developers who b…
numind · Kit
Reads documents as text or images and fills a JSON template with the fields you ask for, or converts the pages to Markdown. A 4B model for teams that process s…
nace-ai · Kit
Scores yes or no, multiple choice and rating questions about a document in one pass, with a probability for every option. An 8B decision model from Nace.AI tha…