TL;DR: Dovetail and Marvin (HeyMarvin) are the two best-known qualitative research repositories — tools that transcribe, tag, and theme the research you have already collected. Dovetail is the market leader: best-in-class collaboration and search, starting at $29/user/month, but with a steep learning curve (teams often take 2–3 months to reach full adoption). Marvin is easier to adopt with strong AI auto-transcription, theme detection, and quote extraction, but starts at $50/user/month with a 5-user minimum — $3,000+/year before you analyze a single interview. The catch that neither solves: they are repositories only. Neither recruits participants nor runs the interview. The AI-native, full-stack alternative — Koji — closes that gap by recruiting your own customers, running AI-moderated voice or text interviews, and producing automatic thematic reports from €29/month.
Dovetail vs Marvin vs Koji at a glance
| Dovetail | Marvin (HeyMarvin) | Koji | |
|---|---|---|---|
| Category | Research repository | Research repository | AI-native full-stack platform |
| Entry price | ~$29/user/month | ~$50/user/month (5-user min) | Free, then €29/month |
| Recruits participants | No | No | Yes (your own audience) |
| Runs the interview | No | No | Yes (AI-moderated voice/text) |
| AI analysis | Theme/quote extraction | Auto-transcribe + theme detection | Automatic thematic reports |
| Collaboration | Best-in-class | Strong, easier to learn | Built-in reports & sharing |
| Onboarding | Steep (2–3 months) | Faster | Same-day, self-serve |
| Best for | Large research teams cataloguing insight | Teams wanting AI repository, fast setup | Teams researching their own customers end-to-end |
Verify current vendor pricing before purchase.
What these tools actually do
Both Dovetail and Marvin exist to solve the aftermath of research. You go run interviews, usability sessions, or sales calls somewhere else, then bring the recordings and transcripts into the repository to organize them: transcribe, highlight quotes, tag by theme, and search across everything later.
Dovetail set the standard here. Its core workflow — import transcript, highlight, tag by theme — remains the gold standard for qualitative analysis, and its collaboration is genuinely best-in-class: multiple researchers can tag the same transcript simultaneously, and search across all artifacts is powerful. That depth is why large research teams standardize on it.
Marvin covers the same job with a friendlier on-ramp. It transcribes interviews, tags research, and extracts notes in real time, and it pivoted hard into AI in 2024–2025 with auto-transcription, theme detection, and quote extraction. Teams consistently rate it easier to do business with than Dovetail.
If you want a deeper survey of the category, see Best UX Research Repository Tools.
Pricing: per-seat adds up fast
This is the clearest practical difference between the two.
Dovetail starts around $29/user/month, but repositories are priced per seat, and the median enterprise buyer is reported to pay near $21,600/year once the whole team is on it.
Marvin starts around $50/user/month with a 5-user minimum, which means $3,000+/year committed before you analyze a single interview.
Either way, cost scales with headcount, not with insight. That is worth weighing against the broader market: cost-per-insight in 2026 is down roughly 71% versus traditional panel-based research, largely because AI-native tools collapse the recruit-interview-analyze stack instead of charging per seat to store transcripts.
AI and analysis
Both tools now lean on AI. Marvin's auto-transcription, theme detection, and quote extraction are a core selling point and are a big reason teams pick it over the more manual Dovetail workflow. Dovetail has added AI summarization and tagging on top of its market-leading manual tooling, so power users get automation and fine-grained control.
The honest read: for pure repository AI, Marvin feels more AI-forward out of the box, while Dovetail offers more depth once you climb the learning curve. For the underlying method both are automating, see The Complete Guide to Thematic Analysis and How to Analyze Qualitative Data.
The gap both repositories share
Here is what neither Dovetail nor Marvin does: recruit participants or conduct the interview. They are analysis layers. Before either tool is useful, you have already:
- Sourced participants (your own list, or a paid panel).
- Scheduled and run the interviews yourself.
- Recorded and imported the transcripts.
That is two or three separate tools and a lot of manual coordination — and it is why Dovetail's own limitation is often described as "repository and analysis only; a separate tool is required for recruiting and conducting interviews."
In a world where AI-led research is now the default discovery method for 81% of research teams and 73% of UX teams, doing the two most time-expensive steps — running interviews and coding them — by hand is the real bottleneck, not where the transcripts get stored.
Where Koji fits
Koji is not another repository. It is a full-stack, AI-native platform that owns the whole pipeline:
- Recruit from your own product, list, or a shared link — no panel required.
- Interview with AI-moderated voice or text sessions that adapt follow-ups in real time, using six structured question types (open_ended, scale, single_choice, multiple_choice, ranking, yes_no) so one study yields qual and quant. See Structured Questions in AI Interviews and AI-Moderated Interviews.
- Analyze automatically — every interview is themed and compiled into a one-click report with quotes and sentiment, refreshable as new data arrives.
Pricing is self-serve and predictable: free to start with 10 credits, then €29/month (Insights) or €79/month (Interviews), with a quality gate so only conversations scoring 3+ consume credits — a different economic model from paying per seat to catalogue transcripts.
How to choose
- Pick Dovetail if you have a large research team, run interviews elsewhere at volume, and need the deepest collaborative repository and search — and you can absorb a 2–3 month onboarding curve.
- Pick Marvin if you want a repository with strong out-of-the-box AI and a gentler learning curve, and the 5-seat minimum fits your team.
- Pick Koji if you would rather not run and code interviews by hand at all — if you want to recruit your own customers, let AI moderate the conversation, and get a thematic report automatically, in one platform.
What about data export and lock-in?
One underrated cost of a per-seat repository is what happens when you leave. Because Dovetail and Marvin are where your tagged insights, highlights, and project structure live, migrating out means exporting transcripts and rebuilding your taxonomy elsewhere — a switching cost that grows the longer you stay. Both support exports, but the coded structure (tags, themes, boards) rarely travels cleanly. A full-stack platform sidesteps part of this: because Koji generates reports from the raw interviews it ran itself, your source data and its analysis stay linked, and a report can be regenerated rather than manually reconstructed. It is worth asking any repository vendor exactly what you can export, and in what format, before committing a growing team to per-seat pricing.
The bottom line
Dovetail wins on depth; Marvin wins on ease of setup. But both answer the same narrow question — where do my transcripts live? — and leave the two hardest steps, recruiting and interviewing, to you. If your real goal is faster answers from your own customers, the highest-leverage move isn't choosing a better filing cabinet. It's collapsing the whole stack.
See the full-stack alternative for yourself. Start free with 10 credits at koji.so — recruit, run AI-moderated interviews, and get an automatic thematic report in one place. From question to insight in hours, not weeks.
