Softwr

Research · head to head

Marvin vs Remesh

Marvin logo

Marvin

Research

Qualitative research repository with AI notetaking, and an MCP server you can query from Claude

From
Free
Rated
-
Remesh logo

Remesh

Research

Live AI-moderated conversations with hundreds of participants at once, analysed in real time

From
On request
Rated
-

The short version

  • Only Marvin has a free tier, so it costs nothing to try first.
  • Each has a real cost: Marvin transcription is inconsistent, with missing punctuation and specifically poor Portuguese accuracy reported, so non-English research teams spend the time cleaning up that the licence was meant to save.; Remesh pricing is not published and third-party estimates put annual contracts in the tens of thousands of dollars, which puts it out of reach for teams that only run occasional qualitative studies and cannot commit to an annual platform fee.
  • They diverge on capability: Marvin covers AI Notetaker, Remesh covers Live moderated sessions.
  • Prices and features above were last checked on 31 August 2026.

Where they differ

Only the attributes on which Marvin and Remesh actually diverge.

Attributes where Marvin and Remesh differ
AttributeMarvinRemesh
Starting priceFreeOn request
Free tierYesNo
PlatformsWebWeb, API

Identical on both: pricing model (quote), user rating (Not yet rated), category (Research).

What each one covers

Drawn from each product's published feature list. An absence here means we hold no record of it - not that the product lacks it.

Only in Marvin

  • AI Notetaker
  • AI Moderated Interviewer
  • Ask AI
  • Deep Research
  • Thematic and emotion analysis
  • Clips and highlight reels
  • Virtual observation room
  • MCP server and API

Only in Remesh

  • Live moderated sessions
  • Collective dialogue
  • Flex asynchronous studies
  • Video research
  • Remy AI assistant
  • On-Demand Recruit
  • Real-time theme analysis
  • Practice sessions

What people use each for

The jobs each tool is most often brought in to do.

Marvin

  • A two-person research team that needs interviews transcribed, tagged and searchable without hiring an analyst or buying an enterprise repository.not Remesh
  • A product organisation where findings are never read, wanting product managers to query past research from Claude or ChatGPT instead of logging into a research tool.not Remesh
  • A team running discovery at volume that wants an AI moderator to conduct unmoderated interviews overnight and have them analysed by morning.not Remesh
  • A researcher who needs to send stakeholders a highlight reel of customer quotes rather than a slide deck summarising them.not Remesh

Remesh

  • A brand team that needs to test three positioning statements with 400 target customers and know by the end of the day which one lands and with whomnot Marvin
  • An HR function running an employee listening session across several thousand staff where anonymity and volume both matternot Marvin
  • A public affairs team needing rapid citizen consultation on a policy option with evidence of how widely each objection is heldnot Marvin
  • An insights team that has to replace six focus groups with one session because fieldwork time has been cut and the decision cannot waitnot Marvin

Where each one falls short

Documented limitations, not opinions. Every one is a constraint you would hit in normal use.

Marvin

  • Transcription is inconsistent, with missing punctuation and specifically poor Portuguese accuracy reported, so non-English research teams spend the time cleaning up that the licence was meant to save.
  • Speaker diarisation errors attribute moderator speech to the participant, and a misattributed quote propagates into themes and into the insight report you show stakeholders, which is a research credibility risk rather than a cosmetic bug.
  • SSO, EU or US data residency, PII redaction and a HIPAA BAA are bundled into a paid Compliance add-on available only from Pro upwards, so the headline Pro quote is never the real price for a regulated or European buyer.
  • Starter includes only two full seats, so a four-researcher team is forced to Pro on seat count alone, and the three-way split between full, collaborator and viewer seats makes cost modelling awkward when no paid tier is priced publicly.
  • Every paid tier is sales-gated with annual commitment typically required, so there is no monthly paid trial and no published seat cost to budget against before entering a sales cycle.
  • Integration coverage has gaps, with no Intercom connector despite repeated requests from teams wanting support conversations in the same repository, and reviewers describe the tags versus labels distinction in the interface as confusing.

Remesh

  • Pricing is not published and third-party estimates put annual contracts in the tens of thousands of dollars, which puts it out of reach for teams that only run occasional qualitative studies and cannot commit to an annual platform fee.
  • The public review picture is inconsistent in a way worth investigating: a small number of positive buyer reviews on G2 sit alongside a poor Trustpilot rating driven mostly by participant complaints, so ask directly how participants are recruited and compensated before running a study on your own customers.
  • Funding has slowed sharply, from a 25 million USD round in 2021 to a 4.25 million USD round in late 2022, so procurement teams at large organisations should expect to run a vendor viability review.
  • The collective dialogue mechanic rewards views that many participants recognise and vote for, which systematically under-weights genuinely novel minority insight, so it is a poor fit for exploratory discovery work where the outlier is the point.
  • The discussion guide builder, especially spreadsheet import, is reported as unreliable, and because sessions are live with recruited participants a tooling failure during setup is expensive rather than merely annoying.

Pricing, plan by plan

Marvin

Free
  • FreeFree
    • 2 full seats and 3 collaborator seats
    • 5 file uploads per month
    • AI notetaking and transcription
  • Starter$undefined/year
    • 2 full seats, 10 collaborator seats, 50 viewer seats
    • Unlimited uploads
    • Thematic analysis
  • Pro$undefined/year
    • 5 full seats, 15 collaborator seats, 100 viewer seats
    • Agentic Ask AI and Deep Research
    • Marvin MCP server and API
  • Enterprise$undefined/year
    • 5 full seats, 15 collaborator seats, unlimited viewer seats
    • Salesforce integration included
    • Dedicated support

Remesh

On request
  • Remesh$undefined/year
    • Live moderated sessions
    • Flex asynchronous studies
    • Video research

Which should you pick?

Choose Marvin if

  • You need ai notetaker.
  • You want to start without paying.
  • You also want ai moderated interviewer.

Choose Remesh if

  • You need live moderated sessions.
  • You work on Web, API.
  • You also want collective dialogue.

Questions people ask

Is Marvin or Remesh better?
Neither clearly leads. Marvin starts at Free and Remesh at On request, and user ratings are close enough to be indistinguishable. Choose on capability and platform support.
Which is cheaper, Marvin or Remesh?
Marvin has a free tier; the other does not. Paid plans start at Free for Marvin and On request for Remesh.
Does Marvin or Remesh run on more platforms?
Marvin runs on Web. Remesh runs on Web, API.
Can I use Marvin for free?
Yes. Marvin has a free tier, so you can try it without paying. Remesh starts at On request.
What is Marvin best used for?
Marvin is most often used for a two-person research team that needs interviews transcribed, tagged and searchable without hiring an analyst or buying an enterprise repository., a product organisation where findings are never read, wanting product managers to query past research from claude or chatgpt instead of logging into a research tool., a team running discovery at volume that wants an ai moderator to conduct unmoderated interviews overnight and have them analysed by morning., a researcher who needs to send stakeholders a highlight reel of customer quotes rather than a slide deck summarising them.. Of those, a two-person research team that needs interviews transcribed, tagged and searchable without hiring an analyst or buying an enterprise repository. and a product organisation where findings are never read, wanting product managers to query past research from claude or chatgpt instead of logging into a research tool. are not what Remesh is typically brought in for.
What can Marvin do that Remesh cannot?
Marvin covers AI Notetaker, AI Moderated Interviewer, Ask AI, Deep Research. Remesh covers Live moderated sessions, Collective dialogue, Flex asynchronous studies, Video research.

Answered from the vendors’ own pages

Marvin: Is there a genuinely free plan?

Yes. Two full seats, three collaborator seats and five file uploads a month at no cost. It is enough to evaluate the analysis quality on real interviews.

Remesh: How many people can join one session?

Sessions are designed for several hundred participants at once, which is the core reason to choose it over a conventional focus group.

Marvin: What does Pro cost?

Not published. Only the Free tier carries a price. Starter, Pro and Enterprise are all quoted.

Remesh: Does it publish pricing?

No. Contracts are quoted annually, typically scoped by session volume, and third-party estimates place them in the tens of thousands of dollars per year.

Marvin: Can I get SSO and EU data residency?

Yes, but as a separately priced Compliance add-on from Pro upwards, not included in any base tier. Price it into the quote from the start.

Remesh: Can it replace a survey?

Not entirely. It produces open-ended language with agreement percentages, which is stronger than a survey for understanding reasoning but weaker for precise population estimates.

Marvin: What is the MCP server for?

It lets Claude, ChatGPT or Copilot query your research repository directly with cited answers, so stakeholders get findings without opening Marvin.

Remesh: Does Remesh recruit participants for me?

Yes, through On-Demand Recruit, typically delivering a panel within about 24 hours, or you can bring your own list.

Share

Related pages

Other head to heads