> Documentation index: https://usertold.ai/llms.txt

# Features: From the First Question to the Shipped Fix

UserTold brings AI-assisted interviews, screen recording review, evidence forensics, and source-linked delivery into one research workflow.

## Ask the next useful question.

Talk with people while the experience is still fresh. Give the interviewer a research goal, then let the conversation follow what the participant actually says and does.

### AI-assisted follow-ups

Follow an answer with a relevant question using your research goal and interviewer instructions. Explore an expectation, a workaround, or a recent decision without scripting every branch.

[Explore ai-assisted follow-ups](/guides/follow-up-questions).

### Observe, then debrief

Let a participant work through a task in silence. Enable pre-debrief analysis for the next Talk step to ask about recent speech, behavior, and page changes.

[Explore observe, then debrief](/features/ai-user-interviewer).

### A flow you can shape

Combine Talk, silent Observe, and scripted Speak steps. Set the goal, instructions, and pacing for the interview you need to run.

[Explore a flow you can shape](/docs/study-runtime).

### Product context for the interviewer

Add product facts and terminology. When more detail is needed, connect an existing read-only knowledge endpoint and enable it for the relevant Talk steps.

[Explore product context for the interviewer](/guides/project-knowledge).

## See the moment. Hear the explanation.

A Finding should take you back to something you can inspect. Review the recording alongside the transcript and the product events that give the conversation context.

### Recorded screen review

Replay the participant’s shared screen alongside their voice. Screen capture is available in supported desktop browsers when the participant grants permission.

[Explore recorded screen review](/guides/in-product-ai-interviews).

### Transcript and timeline

Move between speech, navigation, interactions, and evidence moments. Seek into playback to understand what happened before and after a quote.

[Explore transcript and timeline](/guides/analyze-user-interviews).

### Evidence forensics

Inspect direct quotes, observed facts, behavioral reconstruction, and the claim they support. Page context, preceding actions, and timestamps keep the reasoning connected to its source.

[Explore evidence forensics](/docs/methodology).

### Visible uncertainty

Review confidence, evidence grade, and transcript uncertainty where available. Capture limitations remain visible so missing material does not quietly become a complete story.

[Explore visible uncertainty](/guides/analyze-user-interviews).

## One installation. Different research questions.

Run focused Studies across the same product. Control the invitation and placement so people encounter the right interview in the right context.

### Goals and reusable scripts

Start from practical interview scripts for onboarding, pricing, switching, or feature validation. Adapt the research goal and flow to your product.

[Explore goals and reusable scripts](/guides/research-script-templates).

### Page and language targeting

Use Study visibility to choose where an interview appears and which widget language it matches. Run multiple Studies with one Project snippet.

[Explore page and language targeting](/guides/multiple-interviews-one-website).

### Invitations and direct links

Configure the in-product launcher or share a direct interview link through your own channels. Invite a participant back for a new interview when you need to follow up later.

[Explore invitations and direct links](/guides/follow-up-questions).

### Participant intake and consent

Ask eligibility questions before the interview and present recording disclosures before capture starts. Participants choose whether to begin and grant device permissions.

[Explore participant intake and consent](/docs/studies).

### A localized interview widget

Present the participant experience in English, German, Spanish, French, Japanese, Russian, or Simplified Chinese. Select the widget language for your audience.

[Explore a localized interview widget](/docs/widget).

## Carry the evidence into the fix.

Bring related moments together, review what they support, and send a clear product problem to the team that can act on it.

### Source-linked Findings

Group related Evidence into a Finding. Refine the problem and its supporting context, check conflicting evidence, and review the sources before deciding it is ready.

[Explore source-linked Findings](/guides/prioritize-fixes-with-evidence).

### Linear and GitHub handoff

Explicitly send a reviewed Finding to Linear product intake or GitHub delivery. Keep the user context and linked Evidence attached to the work.

[Explore linear and github handoff](/guides/interviews-to-issues).

### Resolution and recurrence review

Linear completion resolves the linked Evidence. Review later similar Evidence for possible recurrence. GitHub delivery currently provides the handoff without completion synchronization.

[Explore resolution and recurrence review](/guides/track-resolved-evidence).

## Research that people and agents can use.

Use the dashboard for hands-on review, or connect the same research workflow to an agent and your existing development tools.

### MCP, CLI, and API

Set up Projects, operate Studies, inspect interview sources, and work with Findings through documented interfaces. Agents can request focused source windows instead of guessing from a summary.

[Explore mcp, cli, and api](/for-agents).

### Transcript import and artifact access

Import an existing transcript through the API. Access available interview artifacts such as audio, screen recordings, transcripts, and captions for further review.

[Explore transcript import and artifact access](/docs/api).

### Installation verification

Install the Project snippet, verify that the widget can load, and test a complete interview before inviting participants.

[Explore installation verification](/docs/quickstart).

### Managed AI or your own key

Choose managed AI or bring your own OpenAI key. Usage-based pricing keeps the cost connected to recorded interview time.

[Explore managed ai or your own key](/pricing).

## Start with one question worth asking

[Set up your first interview](/docs/quickstart), or [see the complete workflow](/how-it-works).
