01.1
GitHub repo sync
Connect a repository through the official GitHub App and Wingman indexes it into the knowledge graph. Authorize once, revoke anytime.
The AI knows what your product actually does, not what it guesses.
Most tools collect feedback and stop. Wingman runs the whole loop, from raw feedback to shipped code and back.
59 LIVE TODAY·9 SHIPPING SOON·ALL 68 ON THIS PAGE
68/68 FEATURES
Every other tool owns one box and calls it a platform. Wingman owns the whole flight.
Seven stages, one continuous run from feedback to shipped.
Every phase below ends with its full feature ledger — or search all 68 from the bar above.
Connect your product docs and your Git repo. Wingman indexes both into one knowledge graph, then grounds every theme, answer, and PRD in what your product actually does, and what it doesn't. No generic AI fluff. When it talks about your product, it is talking about your product.
UNDERSTANDING SCORE62%
FILES INDEXED1,428
SOURCES LINKED0/5
Connect a repo with the GitHub App. Authorize once, revoke anytime.
Watch the Understanding Score climb as Wingman learns your product: files analyzed, coverage, repos indexed.
Every AI output is grounded in your code and docs. Themes, chat answers, and PRDs cite what is really there.
Upload docs too. PDFs, specs, and design notes, all parsed into the same brain.
Pinpoint the gap between what users keep asking for and what is actually in the code.SHIPPING SOON
When code is cheap, taste is everything. The brain is how Wingman keeps your taste informed.
We read your code so the AI understands your product. Here is exactly what that means.
You connect via the official GitHub App and choose what is indexed. Disconnect or revoke access at any time.
Your index is scoped to your workspace. Retrieval only ever grounds your product's outputs, never another customer's.
Processing runs on frontier LLM providers (Anthropic and OpenAI). Wingman picks the best model for each task; you don't configure it.
01.1
Connect a repository through the official GitHub App and Wingman indexes it into the knowledge graph. Authorize once, revoke anytime.
The AI knows what your product actually does, not what it guesses.
01.2
A live readout of how well Wingman knows your product: files analyzed, coverage, repos indexed. It climbs as the index grows.
You always know how much the brain can be trusted.
01.3
Drop in PDFs, specs, and design notes. Everything is parsed into the same knowledge graph as your code.
Tribal knowledge stops living in someone's drive.
01.4
Every theme, chat answer, and generated PRD retrieves from your code and docs before it writes a word.
No generic AI fluff. When it talks about your product, it is your product.
01.5
Processing runs on frontier LLM providers (Anthropic and OpenAI); Wingman picks the best model for each task automatically.
Frontier quality without a model-picker to babysit.
01.6
Wingman cross-references what users keep requesting against what already exists in the codebase.
Spot the gap between demand and reality before the roadmap does.
Pull feedback from where your customers already are. Wingman ingests it, dedupes the noise, and hands it to the brain. Nothing falls through the cracks because nothing stays in a silo.
Plus spreadsheet upload (CSV/Excel) with AI column mapping. Drop a file, Wingman figures out the columns.
COLLECT · INTAKE STREAM
LIVE
INTAKE
ITEMS CAPTURED · 30D
DEDUPEx41
SPAM FILTEREDx17
A public board where customers post and vote, with vote budgets so it is signal, not a popularity contest.
Your idea board shouldn't be a popularity contest with a subscription fee. Vote budgets keep it honest.
02.1
Connect Intercom, Slack, Zendesk, Jira, and Google Forms. Feedback flows in with submitter, source, and timing mapped automatically.
Meet customers where they already talk. Zero copy-paste.
02.2
Upload any spreadsheet and the AI maps your columns to the right fields, with templates for B2B, B2C, and internal feedback.
A year of backlogged feedback imported before your coffee cools.
02.3
Design a feedback form, publish it, embed it on your site, and collect responses straight into Wingman.
Run targeted research without renting another survey tool.
02.4
Response counts, completion rates, and per-survey breakdowns, tracked from the moment you publish.
Know which questions earn answers, and which get skipped.
02.5
A branded portal where customers post ideas and vote, with vote budgets so enthusiasm can't drown out signal.
An idea board that's evidence, not a popularity contest.
02.6
Review submissions, set statuses, tag, and reply to customers directly from the moderation queue.
Contributors see you listening, in public.
02.7
AI flags duplicate or near-identical feedback pairs with a side-by-side diff and a merge-or-dismiss workflow.
The same request counted once, not five times.
02.8
Microsoft Forms and Google Sheets connectors are in the build queue.
The intake keeps widening without you lifting a finger.
Wingman clusters thousands of raw comments into clear themes, reads the sentiment, flags what is spiking, and lets you ask it anything. The triage you used to do on Monday morning is done before you sit down.
Automatic theme clustering. Feedback grouped by what it is actually about, no manual tagging.
Sentiment and classification. Positive, neutral, or negative; bug vs request vs complaint.
Anomaly detection. A spike or a sentiment shift surfaces before it becomes a fire.
Search across everything. Every theme, every item, one query.
Ask the brain. RAG chat grounded in your feedback and your code.
Feedback is intelligence to interpret, not votes to count.
03.1
AI groups feedback by what it's actually about and writes a summary per theme, with a review-and-save step so you stay in command.
The manual tagging job is gone. The patterns aren't.
03.2
Every item is scored on a five-step scale from very negative to very positive, rolled up per theme, tracked over time.
An instant mood barometer: churn fear or feature gap, at a glance.
03.3
Feedback is auto-labeled bug, feature request, or complaint the moment it lands.
Triage happens before you open the inbox.
03.4
A volume spike or a sentiment shift in any theme gets surfaced the moment it starts moving.
You see the fire when it's a spark.
03.5
Every theme as a card: feedback count, sentiment rollup, priority score, and a drill-down to the raw verbatims behind it.
Your biggest leverage points, visible in one screen.
03.6
A chat grounded in your feedback, your docs, and your code. Ask anything; answers cite what's really there.
Product questions answered with evidence, not vibes.
03.7
One command palette across themes, feedback, and documents, with type facets, relevance ranking, and inline snippets.
Any customer signal found in seconds.
03.8
Filter the feedback table by theme, sentiment, source, language, date, classification, or submitter, with full-text search and highlighting.
"All angry B2B customers asking for export" is one query.
03.9
An AI-composed dashboard of new signals, sentiment trends, and the decisions waiting on you. Ready every morning.
Your 30-second sitrep before the day starts.
03.10
All your products in one grid: feedback volume, sentiment, activity, and priority spread, with 30d/90d/quarter/YTD lenses.
Which product is getting love and which went silent, in one look.
Score themes with the framework you already trust, or your own. Wingman pre-scores impact, effort, and urgency from the feedback itself, so every call has a paper trail.
Five frameworks built in
RANKED QUEUE // Q3 CANDIDATES
AUTO-RESCORE
SCORE = (R × I × C) / E
EFFORT IN SPRINT-WEEKS
AI pre-scoring. Wingman proposes the numbers from the underlying signal; you adjust.
Audit trail. Every score is traceable back to the feedback that earned it.
Defensible decisions, every time. The loudest voice in the room stops winning by default.
04.1
RICE, MoSCoW, Kano, weighted scoring, and fully custom formulas, all built in and switchable per product.
The framework your org already trusts, without the spreadsheet.
04.2
Define your own dimensions: brand fit, team capacity, strategic bets. Set the weights, and score against them.
Company strategy baked into every trade-off.
04.3
Wingman proposes impact, effort, and urgency numbers drawn from the underlying feedback signal. You adjust; it recalculates.
Start from evidence, not a blank scoring grid.
04.4
Every score traces back to the feedback items that earned it, per theme, per framework.
When a stakeholder asks "why this first?", the answer is one click.
04.5
Edit scores directly in the matrix, watch rankings reorder live, and export the result for the deck.
Prioritization meetings end with a decision, not homework.
Every issue tracker is an island. Wingman's Work Board isn't. Issues are born linked to the feedback and themes that created them, the AI breaks epics down for you, and when the work ships, the roadmap updates itself. It is the build stage living inside the same loop as everything else. Not another tool to sync.
Scheduled CSV exports
SSO error copy
Slack digest v2
Webhook retry backoff
Dark mode data tables
Rate limit headers
Kanban that's keyboard-fast. Columns, priorities, labels, assignees.
Issues born from feedback. Link feedback and themes natively; the AI suggests the links.
AI breaks down epics into concrete sub-issues from the linked signal and docs.
“Why this matters”. An AI summary on every issue, grounded in the feedback behind it.
Ship-back loop. Close the work, and the roadmap theme moves to shipped.
Standalone trackers manage work. Wingman manages work that's wired to the customers who asked for it.
Issues born from real customer feedback
Break an epic into sub-issues
“Why does this issue matter?”
Shipping updates the roadmap
Feedback, docs, and work in one place
Grounded in your actual codebase
Everything a tracker does, plus the one thing none of them do: stay connected to the reason the work exists.
One tool, one loop. Rip the sync scripts out.
05.1
Columns, drag-and-drop, and shortcuts for a board that keeps up with you.
Board admin at the speed of thought.
05.2
Four priority levels, custom labels, owners, and deadlines on every issue.
Everything a standalone tracker does, in the same cockpit.
05.3
Per-column titles, colors, descriptions, done-states, and enforced work-in-progress limits.
Workflow discipline the board enforces for you.
05.4
Create issues linked to the feedback and themes that demanded them; the AI suggests the links.
No orphan tickets: every issue knows why it exists.
05.5
Point the AI at an epic and it decomposes it into concrete sub-issues using the linked feedback and docs.
An afternoon of ticket-writing done in a minute.
05.6
Every issue carries an AI summary grounded in the customer feedback behind it.
Engineers see the customer, not just the ticket.
05.7
Threaded discussion on every work item, next to the linked feedback and themes.
Context stays with the work instead of scattering into DMs.
05.8
Close the work and the linked roadmap theme moves to shipped, automatically.
The roadmap updates itself. Rip the sync scripts out.
PRDs, epics, user stories, tech specs, and status updates, generated from your themes and grounded in the brain. Context-aware, editable, regenerate any section. Seconds to a first draft you would actually defend.
PRD: Scheduled CSV exports
DOC-047 // SOURCE: 38 SIGNALS
Generate PRDs, epics, stories, specs, status updates. Not just one doc type.
Streamed live. Watch it plan and write, token by token.
Grounded and editable. Pulls from feedback, docs, and code; edit inline, regenerate a section.
Publish or embed. Share a doc as a public page.
06.1
Generate PRDs, epics, user stories, tech specs, and status updates from your themes, grounded in your feedback, docs, and code.
Seconds to a first draft you'd actually defend.
06.2
Watch the document plan itself and write token by token. No spinner, no black box.
You see the reasoning as it happens, and can stop it early.
06.3
Edit anywhere, or regenerate a single section with an AI rewrite while the rest stays put.
Polish without starting over.
06.4
Every save is kept; roll back to any earlier cut of the document.
Experiment freely: nothing is ever lost.
06.5
Documents link to the themes they serve, and themes list the documents written from them. Both directions, automatically.
Every spec can prove which customer problem it solves.
06.6
Share any document as a public page or embed it outside the app.
Stakeholders read the source, not a stale export.
Drag themes from idea to shipped. Stakeholders see the live view, not last quarter's slide. Publish a public roadmap with your branding.
v2.3CSV import wizard is live06-24 14:07
Idea to planned to in-progress to shipped, drag-and-drop.
Built from your themes, or create items by hand.
Public roadmap. Branded, embeddable, always current.
07.1
Move themes from idea to planned to in-progress to shipped on a visual board, built from your themes or by hand.
The roadmap is a control surface, not a slide.
07.2
Publish the roadmap as a customer-facing page or embeddable widget, in your branding, always current.
Stakeholders see the live view, not last quarter's deck.
07.3
Choose per column what goes public automatically and what stays internal.
Transparency on your terms.
07.4
A draft-to-publish workflow with sections for features, improvements, and fixes, plus AI summaries of what shipped.
Release comms written once, published everywhere.
07.5
Every release gets a shareable public URL, and the whole log can be embedded on your site.
Your ship record becomes marketing that writes itself.
Collecting feedback is the easy half. Wingman does the half nobody else finishes: it goes back to the people who spoke. When feedback is vague, it spots it, drafts the clarifying question, and emails the contributor, automatically or on your sign-off. Their reply lands back in Wingman, threaded to the original signal. The conversation lives next to the feedback, not buried in someone's inbox.
Spot vague feedback and ask the right question. AI-drafted, 1 to 3 questions that actually change the triage.
Send automatically or on approval. Clarification, status-change, thank-you, and survey follow-up rules, with your own timing and templates.
Replies come back in, threaded to the feedback item, with unread tracking, so the loop is a conversation, not a broadcast.
Tell contributors when you ship. The people who asked hear that it is done.
Auto-fold each reply back into the theme: re-clustering and enrichment from the answer.SHIPPING SOON
THREAD FDB-2317
CLOSE-LOOP
“We export usage CSVs by hand every Monday. If Wingman could schedule those exports we’d save half a morning, every week.”
23 REPORTERS NOTIFIED2 REPLIED
Draft and publish a branded, embeddable changelog. AI summarizes what shipped.
Speed without signal crashes planes. Closing the loop is how the signal gets back to you.
08.1
When feedback is too thin to act on, Wingman spots it and drafts one to three clarifying questions that would actually change the triage.
"It's slow" becomes a reproducible report, without you chasing it.
08.2
Clarification, status-change, thank-you, and survey follow-up emails, sent automatically or on your approval, with your timing and templates.
Every contributor hears back. You approve, not compose.
08.3
Customer replies land back in Wingman threaded to the original feedback item, flagged until read.
The loop is a conversation, not a broadcast.
08.4
When work linked to feedback ships, Wingman can tell the exact people who asked for it.
The fastest trust-builder in product: "you asked, we built it."
08.5
Each answer folds back into theme analysis, re-clustering and enriching the signal automatically.
Follow-ups make the brain smarter, not just the inbox fuller.
The last question every roadmap dodges. Wingman is building the answer: trace a shipped feature back to the feedback that demanded it, and forward to whether it moved retention and revenue.
Impact tracker. What got built from which customer voice.SHIPPING SOON
ROI signals. Did the shipped work move the metrics.SHIPPING SOON
A loop you can't measure is a guess with extra steps. This is the part that ends the guessing.
MENTIONS/WK
4141
SENTIMENT
2.12.1
ADOPTION
12%12%
09.1
Trace the full chain from feedback → theme → score → document → work item → shipped, and export it as a report.
"This feature came from 47 customer requests." With receipts.
09.2
Link shipped features back to the customer voices that demanded them and track what moved after release.
Quantified impact per feature, not anecdotes.
09.3
Estimate revenue impact before building, from feedback volume, sentiment, and segment size.
Bet on features with the odds in view.
09.4
Cross-theme trend analysis, contributor analytics, and sentiment time-series in one place.
Spot opportunities before they spike.
09.5
Aggregate what customers love and hate about you versus the alternatives they mention.
Positioning informed by verbatims, not hunches.
09.6
A weekly digest of sentiment, top themes, shipped items, and upcoming priorities, composed automatically.
The leadership update writes itself.
Run B2B, B2C, and internal products side by side — feedback, themes, scores, and roadmaps isolated per product. Roles for the team, real-time sync, your branding on everything public, and your data exportable any day you want it.
10.1
Run B2B, B2C, and internal products side by side: feedback, themes, scores, and roadmaps isolated per product.
A whole portfolio without cross-contamination.
10.2
Invite teammates as admins, members, or viewers with workspace-level permissions.
Everyone gets exactly the access they need.
10.3
Every change propagates to every teammate instantly. No refresh, no stale boards.
One board, one truth, zero "did you reload?"
10.4
Set your logo and colors once; public roadmaps, changelogs, and idea boards inherit them.
Public artifacts that look like you, not like your vendor.
10.5
Email and in-app alerts for high-priority feedback, mentions, and shipped work, filtered by relevance.
Aware without the notification firehose.
10.6
Pull feedback, themes, scores, or roadmap data out as CSV or JSON whenever you want.
Your data stays yours. No ransom at the exit.
10.7
A first-run wizard walks you from product setup to first import to first themes, with contextual guidance in every empty state.
From signup to first insight in one sitting.
10.8
Upgrade, downgrade, and track usage (API calls, surveys, emails) without talking to sales.
The plan follows your needs, not a sales cycle.
10.9
A documented REST API plus Zapier recipes for wiring Wingman into anything else.
If it has a webhook, it can join the loop.
Stop stitching a feedback tool to a roadmap tool to a Jira board to a doc editor. Run the loop in one place, from the first piece of feedback to the shipped code and back to the customer who asked.
Closed alpha. Limited spots.
Still comparing tools? All 68 features are searchable on this page.
You fly the product. We cover your six.
For most product teams, yes. The Work Board handles columns, priorities, labels, assignees, and AI epic breakdown. Unlike a standalone tracker, every issue is linked to the customer feedback that created it, and shipping it updates your roadmap automatically.
Wingman indexes your repo into a knowledge graph through the official GitHub App. From then on, every theme, chat answer, and generated PRD is grounded in what your product actually does, not generic assumptions. You can revoke access at any time, and your index only ever grounds your own workspace.
You connect through the official GitHub App and choose what is indexed. Your index is scoped to your workspace and only grounds your own product's AI outputs. Processing runs on frontier LLM providers (Anthropic and OpenAI), and Wingman picks the best model for each task.
Yes. When feedback is vague, Wingman drafts a clarifying question and emails the contributor, automatically or on your approval, and their reply comes back threaded to the original feedback. When you ship, it can tell the people who asked. Automatically folding replies back into theme analysis is shipping soon.
Intercom, Slack, Zendesk, Jira, and Google Forms today, plus CSV/Excel upload with AI column mapping, built-in surveys, and a public ideas portal. Microsoft Forms and Google Sheets are on the way.
RICE, MoSCoW, Kano, weighted scoring, and fully custom formulas, with AI pre-scoring of impact, effort, and urgency drawn from the underlying feedback.
Wingman is in closed alpha. Request early access from any button on this page. Spots are limited.