Overview

QRetro is a completely free retrospective tool with no seat or team caps, built by solo developer Roman Ermilov and shipping since January 2023. It runs a fixed seven-stage session — optional health check, optional icebreaker game, cards, grouping, voting, discussion, results — and closes with an OpenAI-generated summary, a ROTI score, suggested next steps and an analysis layer that clusters the cards by topic, sentiment and urgency. An official, open-source MCP server exposes 29 read/write tools so Claude, Cursor or VS Code can query boards, health trends and actions directly.

Best for: Small agile teams and independent squads that want a genuinely well-equipped retro — 52 templates, health checks, ROTI, timer, in-board surveys, icebreaker games and MCP access — at zero cost, and can accept a solo-maintainer vendor with no SSO, no PDF or spreadsheet export and no certifications.

How QRetro describes itself

QRetro pitches itself as "an operating system for continuous workflow improvement", with the homepage line "Manage team health, not sticky notes." The argument is that the board is the least interesting part of a retrospective and that what matters is what survives the meeting.

"Most retrospective tools are just digital boards where ideas die when the tab closes." "Meetings end. Actions remain."

Two things are foregrounded unusually hard for a free tool. Health is a first-class output rather than a template — the February 2026 release notes describe the goal as moving QRetro from "just a discussion board" to "a regular team diagnostic tool." And the analysis engine is sold as something that works quietly rather than as a chatbot: "The engine runs in the background, autonomously finding hidden patterns without intrusive chatbots." Where other vendors ship an in-product assistant, QRetro ships an MCP server and tells you to "bring your own model" — "one address, nothing to paste." The voice is engineer-direct and anti-bloat, and the footer signs off "Made for teams that talk straight."

Our take

QRetro is the best-equipped tool here that charges nothing at all — no paid tier, no seat caps. The seven-stage flow covers what facilitators reach for, and it adds things zero-cost tools almost never have: a radar-charted health check that trends across retros, ROTI with a morale trend line, up to five in-board surveys, presentation mode, and four built-in games as a switchable icebreaker stage. 52 templates is competitive with paid products, and guests join by link with no account. Be clear about the comparison, though: several paid tools have free tiers that are better equipped than this, Parabol's among them.

Our August entry undersold the analysis layer, and it is the most interesting thing here. Cards are scored for sentiment, category and urgency; the Priority Matrix plots team votes against that urgency score and calls out the top-left quadrant — high risk, few votes — as the thing the team is avoiding. The Perception Mirror goes further and audits the team's own grouping, naming the groups that hold three or more distinct topics and the topics nobody grouped at all. That is a genuinely useful facilitator check, and no other free tool here does it. The MCP server is the reporting surface: official, GA at v1.3.0, one URL with OAuth, 29 read/write/delete tools and two analysis prompts, with the source public on GitHub.

The limits are structural, not a missing release. This is one developer's product — no support team, no SLA, no published entity, no certifications, no named hosting location — which rules it out of enterprise procurement however capable the board is. Export is thinner than the analysis deserves: you can copy the summary as Markdown, email it to the team, push it to Slack or Telegram, and download your own account as JSON, but there is no PDF and no CSV, so the numbers stay inside the product. The flow also assumes everyone is in the session live — no async mode, no recurrence, no scheduling.

Best for small teams and independent squads who want a serious, engaging retro with health tracking, real post-retro analysis and AI-assistant access at zero cost; not for regulated orgs, distributed async teams, or anyone who needs the retro record as a spreadsheet.

Score breakdown

Ease of Use8.5

Retro Toolkit7.0

Value9.0

Fun Factor7.0

AI & Insights7.5

Integrations5.0

Open Platform4.5

Enterprise-grade2.0

Pros

+ Genuinely free with no seat caps, team caps or paywalled features — a 24-person org across three teams pays nothing, and there is no paid tier to be upsold to
+ Retro toolkit is deep for a free product: 52 templates, editable columns, private-mode submission, a timer, drag-and-drop grouping with auto-named groups, presentation mode, up to five in-board surveys, action items with owners and due dates, ROTI and a radar-charted health check stage
+ The analysis layer is the surprise — every card gets a system-assigned sentiment, category and urgency score, and the Perception Mirror compares the system's own topic clusters against the groups the team made and names the ones that quietly hold three different problems
+ Unusually strong engagement layer for a product that charges nothing — four built-in games (Draw & Guess, Sprint in one GIF, Hangman, Decoded) usable as a switchable icebreaker stage, plus GIF cards, emoji reactions and live cursors. TeleRetro and the whiteboard tools still go further, mainly on music and ambient audio
+ Official MCP server at a single address (mcp.qretro.com) with 29 read, write and delete tools, two analysis prompts and OAuth sign-in — verified live at v1.3.0, and the server source is public on GitHub
+ Guest access with no account or password, and guest contributions persist permanently rather than expiring after a day
+ Visibly active for a solo project: 14 dated changelog entries in 2026 (MCP, OAuth, poker over MCP, two new games), legal pages rewritten in late July 2026 and a site rebuild on 15 August 2026

Cons

− Built and maintained by a single developer (Roman Ermilov) with no support team, SLA or published company entity — the enterprise-grade cap for founder-only vendors applies regardless of how good the product is
− No enterprise controls: no SSO or SAML, no SCIM, no audit logs, no SOC 2 or ISO 27001, and no published hosting location or data-residency option
− Export is thin — a Markdown copy of the summary, an email to the team, a Slack or Telegram recap, and a JSON dump of your own account data. There is no PDF, CSV or Excel, so there is nothing to hand a stakeholder and nothing to drop into a spreadsheet
− Integration set is narrow: Jira, Linear, Slack and Telegram only — nothing for Teams, Confluence, GitHub, GitLab, Azure DevOps or any other tracker
− No async or distributed-team support — the seven-stage flow assumes a live session, and there are no recurring or scheduled retros
− Vote allowance is computed automatically (cards + 3, capped at 10) rather than set by the facilitator, the timer maxes out at 30 minutes, and the stage order itself cannot be customised
− No working-agreements artifact that survives between retros — what the vendor calls an 'agreement' is an action item on a board
− No cross-team or org-level reporting: the Analytics section of the dashboard is still labelled Coming soon, so health and action data stays inside a single team's boards

Features

Retrospectives

Templates· 52 templates including Mad Sad Glad, Sailboat, 4Ls, Starfish, Start Stop Continue, Lean Coffee, Six Thinking Hats, Fishbone, SWOT, pre-mortem and seasonal formatsCustom templates· Columns can be added, renamed, recoloured and reordered when the board is created, and a previous board's layout can be reused; there is no saved named template libraryAnonymous input· Private mode hides cards until the stage completes and vote counts can be hidden; card authors can optionally be shownPrivate/public retrosCustom process flow· The health-check and icebreaker stages can be toggled on or off, but the seven-stage flow itself is fixed

Health checks

Health checks· Optional opening stage — participants rate 6 aspects (communication, task clarity, manager support, vision, blockers, motivation) on a 1-10 scaleTeam radars· Health-check results render as a radar chartMood / sentiment check· Team health dynamics tracked across successive retrospectivesPsychological safety· Safety Check retrospective template is available, but there is no tracked psychological-safety metric

Facilitation

Meeting timerAction itemsLink actions to ideasPresentation mode· Presentation mode hides personal card controls so the board can be projected to the roomAgenda planner· Fixed seven-stage flow: health check, icebreaker, cards, grouping, voting, discussion, resultsDrag-and-drop grouping· Drag-and-drop grouping of similar cards with automatic group-name suggestions

Voting

Independent dot votingVote on ideas inside groups

Engagement

ROTI feedback· ROTI (Return on Time Invested) captured at the end of each session and included in the shared recapPolling· Up to five in-board surveys per retro, each with 2-10 options; the system can draft the question and the options from a plain-text descriptionIcebreakers· Optional icebreaker stage with four built-in games, switchable mid-roundIcebreaker games· Four built-in games in the optional icebreaker stageGIFsPlanning poker· Planning poker with Fibonacci, T-shirt and custom decks, votes hidden until reveal, and Jira/Linear sprint importTeam kudos· Kudos and Appreciation retro templates rather than a standalone recognition feature

AI

AI grouping / clustering· The system derives its own topic clusters from the cards and shows them as the Priority Matrix and the Perception Mirror, which flags team-made groups that hold three or more different topics; grouping on the board itself stays manualAI template suggestions· Enter a theme and a format and the system generates the retro columns, filed as an Auto-generated templateAI action items· Suggested next steps derived from the discussion, each promoted or rejected individually and carrying the cluster it came fromAI summaryAI sentiment / heatmaps· Every card and cluster carries a system-assigned sentiment, category and urgency score, plotted on the Priority Matrix and readable over MCPAI retro / health-check insights· Board summaries, topic clusters, health Key Signals with anomaly detection and an alignment score, plus ROTI and health trends across previous retros — all queryable over MCPAI discussion prompts· Two built-in MCP prompts for retrospective analysis and health-trend comparison

Reporting

Action tracker dashboard· Action items aggregated in a Tasks view across boardsExport formats· Markdown, Email, JSON

Enterprise

GDPR compliant· Privacy policy maps each processing purpose to a GDPR Art. 6 legal basis, with a published DPA, a dated sub-processor list, self-service account data export and account deletion; no certification and no named hosting locationFacilitator-only roleGuest / observer roles· Guests join by shared link with no account or password; their cards, comments and action items now persist on the board permanentlyMeeting historyEmail domain restrictions· A domain-access toggle lets anyone with a company email address join the team without an invite, and emails the owner each time; it opens access rather than restricting who may be invitedMCP server

Integrations

Jira· Import past sprints for analysis and planning poker; generate tickets from retrospective outcomesLinear· Create issues directly from retrospective outcomes; sprint import for planning pokerSlack· Share board links to a channel and post the finished recap (participants, cards, ROTI, summary, agreements and top-voted items)

Compare QRetro against

Sources

Last reviewed: September 1, 2026