How to use Feedbaq

Getting started

Create an account with your email or with Google. There is no password: you sign in with a one-time code. Your first sign-in sets up your workspace, and from there you can start a study three ways: from a template, from a blank study, or by asking your AI agent to build one. Building and previewing studies are free. You subscribe, for $49 a month, when you publish.

Building a study

A study is a welcome page, your questions and a thank-you page. For each question, you choose how people answer: by voice, by a narrated screen recording, in text or by picking from a list. You can also let a question take more than one format, so people who cannot talk right now can type instead.

Turn on AI follow-up questions where you want to hear more. The AI then asks about what each person said, up to 3 times, one question at a time. You can give the AI a line of guidance for each question, such as “ask what they did instead”, and decide whether people may skip the follow-ups.

Add instructions or an image under a question, make a question optional, let a choice question take several answers, and drag questions into the order you want. The builder saves as you go.

Screen recordings

A screen question asks people to share a window or their whole screen and talk while they work. You can add a link to the page you want them to try, and it opens before the recording starts. Screen sharing needs a computer, so for people on a phone or a tablet you choose what happens: offer another answer format, skip the question, or ask them to open the study on a computer.

Templates and agent prompts

Feedbaq comes with study templates and with prompts you can send to your AI agent. There are templates for product feedback, concept tests, churn interviews, customer discovery and more, and each one has a matching prompt that asks your agent to build that study for your product. Other prompts help your agent sharpen a draft, or start a study from your own analytics, call notes or support tickets.

The look and the link

Add your logo, choose a font and set the colours, and the study looks like yours. Preview runs the real participant flow, so you can answer your own study by voice or screen recording and see the transcript and the AI follow-up. Nothing from the preview is saved.

Every study gets a link of the form feedbaq.to/your-workspace/your-study, and you can change the last part to something readable in the Share dialog.

Publishing and versions

Publishing makes the study link live and freezes the questions as version 1. You can keep editing afterwards; participants see the published version until you publish again. Every response remembers the version it answered, so your results stay consistent when the questions change.

You can set a study to stop after a number of responses, close it and reopen it later, or duplicate it to start a new draft from the same questions.

What participants see

Participants open the link in their browser, with no account and no app, and answer one question at a time. The browser asks for the microphone at the first voice question, and if someone declines, the question switches to typing. If they reload the page in the same tab, they continue where they left off. People can answer in their own language: transcription detects the language they speak, while the study screens themselves are in English.

Microphone permissions

If a participant blocked the microphone, the voice question switches to typing and shows how to turn the microphone back on. The steps depend on the browser.

• ChromeOpen the Chrome menu at the top right and choose Settings. Go to Privacy and security, then Site settings, then Microphone. Under Not allowed, select the study site and set it to Allow, then reload the page.
• EdgeClick the site information icon at the left of the address bar, open Permissions for this site, set Microphone to Allow, then reload the page.
• Safari on iPhoneTap the aA button in the address bar, open Website Settings, set Microphone to Allow, then reload.
• Safari on MacOpen the Safari menu, choose Settings for This Website, set Microphone to Allow, then reload.
• FirefoxClick the permissions icon at the left of the address bar, clear the blocked microphone entry, then reload.

Results and analysis

Every voice answer and screen recording comes with a transcript, timed word by word, so you can click any word and play the recording from that moment. Open answers get labels by topic, such as Feature Request or Cancellation Reason, and the Results tab counts those labels across participants and shows the answers worth highlighting. You can group and filter responses, add or remove labels, and export a study to CSV.

To turn many answers into a decision, work in this order:

1. Start from the counts. Open Results and look at the biggest labels. They show what is common, and what is one loud voice.
2. Read and hear the quotes behind them. Play the moments, and highlight the answers that say it best.
3. Ask your AI agent for findings with evidence. It reads every response, so ask for the most common reasons, each with how many people gave it, two quotes and the links. The AI analysis tab has prompts to start from.

Ask for counts and quotes rather than a summary. A summary reads well but hides how many people said something, and you cannot check it without rereading everything. With a count and a linked quote behind every finding, you can trust the result and show it to your team.

Team and roles

Invite colleagues under Settings → Workspace. Every member can build, publish and close studies, read responses, and delete studies and responses. Admins can also invite people, change roles, manage billing and set the monthly response cap. The owner can also delete the workspace.

Plan and billing

One plan at $49 a month with every feature. Connecting your AI agent, building and previewing studies are free; you subscribe when you publish. Your usage, the monthly response cap and your invoices are under Settings → Usage & billing. See pricing.

Limits

Good to know before you plan a study:

• Voice answers and screen recordings run up to 10 minutes each. Screen recordings capture the screen and the voice, not the camera and not system sound, and they need a computer.
• AI follow-ups: up to 3 per question.
• Questions run in the order you set, so every participant gets the same study and the answers compare cleanly.
• The study screens and the app are in English. Participants can speak any language.
• Feedbaq collects answers from the people you invite; it does not recruit participants or send out the link for you.
• Take results anywhere as CSV, or let your AI agent read them directly.

Your data

Recordings, transcripts, typed answers and choices stay as long as the workspace exists. Deleting a response removes its recordings and transcripts at once. The security page explains storage, AI processing, retention and access.