How it works
It is just the conversation
No dashboard to learn first, no forms to fill in. Say what you want to find out, and check in when you want to know where it got to.
I need to talk to 15 people who manage payroll at companies with 50 to 500 employees in the US. What would that cost?
- Terac
Sent that over for a confirmed price, since a panel that specific is worth pricing by hand. It comes back out of band and I will pick it up the moment it lands.
terac_request_feasibility
Good. Set it up as a 20-minute interview about how they choose payroll software.
- Terac
Built the draft: targeting on country and company size, a screener that separates people who actually own the decision from people who only use the tool, and a 20-minute interview task. It comes back priced at an estimated $45 per participant, so $675 for 15, plus the platform fee, and the confirmed price replaces that when it arrives. Nothing is live and nothing is charged. Here it is to review.
terac_create_opportunity
Looks right. Launch it.
- Terac
Live and recruiting. I will check progress when you ask.
terac_launch_draft_opportunity
Any results yet?
- Terac
9 of 15 completed, 6 awaiting your review. Two look thin on the decision-ownership question. Want me to walk you through them?
terac_get_submissions
An illustration of the flow, not a recorded session. Wording varies; the tools and the order they run in do not.
What you can do
A whole research study, from one conversation
Find out what it costs, first
Describe the audience and the draft comes back priced per participant, before anything is committed. When the number is worth getting exactly right, a person on our team prices it by hand and tells you whether the panel is reachable at all, which is the part that usually sinks a study late.
Recruit people who actually qualify
Target on verified profile attributes, then screen with questions written from what the task really requires. An AI moderator interviews applicants to re-check what they claimed.
Run interviews, tests, and surveys
Point people at your own survey or prototype, or have Terac run the task. Either way the work comes back attached to the person who did it.
Review the work and pay on approval
Read submissions, approve the good ones and reject the rest with a reason. Approving is what pays the expert, so quality stays yours to control.
Install
Three steps, about a minute
- 01
Open the app
Find Terac in ChatGPT's app directory, or follow the link below.
- 02
Connect your account
Sign in with the account you already use. No organization yet? You can create one right here and you will be brought back to finish connecting.
- 03
Pick your organization
Terac asks which organization the app acts for. Everything it does afterwards is scoped to it, and you can see all of it in the dashboard.
Using Claude, Cursor, or something else? The same tools work there too.
FAQ
Question not answered? Email us at support@terac.com or book a meeting with our team.
Book a meetingIt lets you recruit and run research with real people without leaving the conversation. Describe who you need to hear from, and Terac prices it, finds and screens matching people, collects their answers, and pays them. You review the work and approve it, all from chat.
No. Install the app, connect, and you can create your organization during the connection itself. You will be brought straight back to finish connecting.
No. Building a study and getting it priced are both free, and nothing recruits anyone or charges you until the study is launched. You can go all the way to a fully built, fully priced study and decide not to run it.
Terac's own verified panel, not anonymous crowdworkers. Every expert is identity-verified through attestations such as government ID, work email and LinkedIn, and an AI moderator screens and interviews applicants before their work reaches you.
Only from money you have already added. Launching is the one action that commits funds, and the app asks you to review the priced draft and approve it before it launches. Your organization's balance is the hard limit either way: a launch that exceeds it fails rather than billing you. You can see and control all of it in the Terac dashboard.
Recruitment is asynchronous. Simple audiences can start returning submissions within hours; niche or hard-to-reach ones take longer. The app can check progress whenever you ask, and the dashboard shows live fill against your target.
Yes. The same tools are available to Claude Code, Claude Desktop, Cursor, VS Code, and any client that speaks the Model Context Protocol, and directly over the REST API.