Exporting Research Data from Koji: CSV, JSON, and Transcript Access

The short answer: Koji gives you complete ownership of your research data. Export participant data and AI-generated summaries as CSV, download raw structured analysis as JSON, access individual transcripts, and pipe data to external systems via the API, all with a few clicks or a single API call.

Your research data is valuable. The insights you collect through Koji interviews should not live only inside a platform — they need to flow into your product roadmaps, CRM systems, research repositories, stakeholder reports, and analytics pipelines. That is why Koji is built for data portability from the ground up.

This guide covers every way to get your data out of Koji, when to use each format, and how to build automated pipelines that keep your research connected to the rest of your stack.

---

Why Data Export Matters in Modern Research

Research that lives in a silo does not drive decisions. The most effective research teams connect their interview data to:

• Product roadmaps — tagging insights to features and epics
• CRM systems — enriching customer profiles with interview notes
• Research repositories — building a searchable archive of findings
• BI tools — creating dashboards that track sentiment and themes over time
• Stakeholder reports — presenting findings in formats teams can consume

With tools like Koji, all of this is possible without manual copy-pasting. Koji is designed so that data flows to wherever it is needed, in the format that each system requires.

---

Export Format 1: CSV

Best for: Spreadsheet analysis, CRM uploads, research operations reporting

CSV export gives you a flat-file view of your participant data, ideal for quick analysis in Excel, Google Sheets, or Airtable, and for uploading to CRM systems or research repositories.

What is included in the CSV

When you export a study to CSV, Koji includes:

• Participant metadata: display name, email (if collected via lead form), source (organic, custom link, CSV import), started and completed timestamps
• Session data: session duration in seconds, message count, interaction mode (voice or text)
• AI summary per participant: the goal-aligned summary of what that participant said
• Quality score: the 1–5 quality score assigned by Koji's analysis engine
• Theme tags: AI-generated tags identifying the main topics discussed
• Intake form data: any fields collected via your lead collection form (name, company, role, etc.)
• Structured question responses: for studies using structured questions, each question gets its own column with the participant's response

How to export CSV

1. Open your study in Koji
2. Open the Interviews page and its Self-recruit tab
3. Click the Export button in the top-right corner
4. Select CSV as your format
5. Click Download

The file downloads immediately — no waiting, no email link. For large studies (100+ participants), export typically completes within a few seconds.

Tips for CSV export

• Filter before exporting: use the quality score filter to export only high-quality interviews (score 3+) before downloading
• Re-run analysis first: if you recently refreshed your report, re-export CSV to get updated theme tags and summaries
• Structured question columns: studies using structured questions (scale, single_choice, yes_no, etc.) will have quantitative columns that are ideal for pivot tables and distribution analysis

---

Export Format 2: JSON

Best for: Developer workflows, research repositories, custom analysis pipelines, AI tools

JSON export provides the full structured analysis for each participant interview — the same data Koji uses internally to generate reports. It is richer than CSV and includes nested data that does not flatten well into rows and columns.

What is included in the JSON export

Each participant entry in the JSON export includes:

• All CSV fields (metadata, session data, summary, scores)
• Full question-answer mapping: each key question from your research brief paired with the extracted answer and confidence level
• Structured answers: for structured questions, the full StructuredAnswer object including quantitative value, qualitative answer, follow-up insights, and confidence
• Pain points, feature requests, and positive highlights extracted by the AI
• Notable quotes: direct quotes from the transcript worth highlighting
• Topic insights map: for each topic in your research brief, what the participant said
• Evaluation: whether the participant answered key questions, stayed on topic, and provided depth
• Full transcript: all messages in the conversation, including AI follow-up questions

How to export JSON

JSON export is available through the Koji API:

GET /api/v1/projects/{project_id}/export?format=json

Include your API key in the Authorization: Bearer {api_key} header. The response is a JSON array of participant analysis objects. See the API Authentication guide for how to generate and manage API keys.

For programmatic access, pair the JSON export with a server job that polls the read endpoint, or with the embed widget's browser events (koji:interview_completed), to keep a pipeline current. Outbound webhooks are not available yet.

---

Export Format 3: Transcript Access

Best for: Qualitative analysis, research repositories, compliance, stakeholder deep-dives

Individual transcripts let you read the full conversation between Koji's AI interviewer and each participant. Transcripts are accessible directly in the Koji interface and can be copied for use in analysis tools.

Accessing transcripts in Koji

1. Open your study's Interviews page
2. Click on any participant row to open the Analysis Drawer
3. Click the Transcript tab to view the full conversation
4. Use the copy button to copy the entire transcript to your clipboard

Transcripts show both the AI's questions and follow-up probes alongside the participant's responses. This gives context that summary data cannot fully capture — tone, hesitation, unexpected tangents, and specific language patterns.

Using transcripts with AI tools

Many research teams paste transcripts into AI assistants for custom analysis. Koji's transcripts are clean, well-formatted, and export-ready. Note that if your research involves personal data, review your privacy policy before sharing transcripts with third-party AI services. If you use Claude or another MCP assistant, the Koji connector reads transcripts and study data directly, so nothing has to be copied and pasted; inside Koji, every claim in the report links to the quote behind it.

---

Export Format 4: API Polling and Embed Events for Automated Pipelines

Best for: Automated research operations, CRM enrichment, keeping your own tools in sync

For teams that want research data flowing automatically into other systems, Koji's REST API enables pipelines that require no manual export steps: a server job polls the read endpoint for new interviews, and the embed widget emits browser events you can act on.

Polling and embed events

Outbound webhooks are not available yet. Two mechanisms cover the same jobs:

• Poll the read endpoint from a server job (or from Zapier, n8n or Make) to pick up each interview's full analysis once it completes
• Embed events: the embed widget emits browser postMessage events (koji:ready, koji:interview_started, koji:interview_completed) that your own product can react to

With a polling job in place, you can automatically:

• Create a Notion or Confluence page for each new interview
• Post a Slack summary to your research channel
• Update a CRM record with the participant's key insights
• Trigger analysis in a custom data pipeline
• Archive transcripts to a research repository

Headless API for embedded research

If you have embedded Koji interviews in your product using the Headless API, all data is available through the same export mechanisms. Participants from embedded interviews appear on the Interviews page and are included in all exports.

For more on API-driven exports, see the Headless API Overview and Webhook Setup guides.

---

A Complete Research Ops Export Workflow

Here is a common workflow for research teams that need both speed and depth:

1. During fieldwork: pin the public report link in Slack, or have a job poll the API and post summaries, keeping your team informed as responses come in
2. After fieldwork: export CSV to review participant-level quality scores and filter out low-quality sessions
3. For synthesis: export JSON or access individual transcripts for deep analysis in your research repository
4. For reporting: use Koji's built-in reports (with optional publishing) for stakeholder distribution — no export needed

This layered approach means you always have the right level of data for each use case, without drowning in raw data you do not need.

---

Data Ownership and Privacy

Koji does not sell or share your research data. All interview data, transcripts, and AI analysis belong to your account. You can delete studies and all associated data at any time.

For enterprise teams with specific data residency or retention requirements, contact the Koji team to discuss enterprise data handling options.

---

Related Resources

• Structured Questions Guide — add quantitative columns to your exports with scale, choice, and ranking questions
• Webhook Setup — keep pipelines current with API polling and embed events
• Headless API Overview — programmatic interview management and data access
• Managing Research Participants — the Interviews page in depth
• Research Automation: How to Build Real-Time Research Pipelines — end-to-end automation workflows
• Generating Research Reports — Koji's built-in report system for stakeholder sharing

Further reading on the blog

• Agile User Research: How to Run Continuous Research in Sprint Cycles (2026) — Most teams know they should do user research every sprint. Almost none actually do. Here's the practical playbook for integrating continuous
• AI Agents for User Research in 2026: How Autonomous Research Is Reshaping Customer Insight — AI agents are taking over user research in 2026 — moderating interviews, synthesizing themes, and producing insight reports in hours. The fu
• Best AI Market Research Tools in 2026: The Complete Buyer's Guide — AI has fundamentally changed market research. This guide compares the leading AI market research platforms—from AI-native interview tools li

<!-- further-reading:blog -->