theyasked.me

Compare · Facts about Canny checked 2026-10-22

Why AI product teams pick TheyAsked over Canny

Canny is a mature feedback suite with AI that pulls requests out of calls and tickets. TheyAsked is built for teams where an agent does the building: it tells Claude, Cursor or your own agent what customers asked for, who asked and what they pay. Here is how the two compare, checked on Canny's own pages.

Start free · Move from Canny

Short answer

Pick Canny if your feedback lives in sales calls, support tickets and review sites and you want AI to extract it for you, or you need a large catalog of CRM, support and project tool integrations. Pick TheyAsked if your team works in Claude, Cursor or Claude Code and wants the agent to ask "what should we build next and who is waiting for it" and get a ranked, revenue-weighted answer in one call, with read-only and scoped access you can hand to an agent safely, at a flat $29 a month that does not grow with your voters.

Why AI product managers and engineers pick TheyAsked

  1. Revenue ranking is a tool your agent calls

    TheyAsked ranks requests by unique voters, the MRR of the companies voting, momentum (votes in the last 14 days against the 14 before) and opportunity. Your agent calls top_requests with a plan or company filter and gets the ranked list back. Canny shows MRR and opportunity value too, but its MCP leaves the ranking to the model through views and segments.

  2. Who asked, in one call

    who_asked_for returns the plan mix with revenue, the companies with their MRR, platforms, importance and the voters' own notes for one request. With most tools that takes a voter list, then a lookup per user, then aggregating by hand. One call keeps the agent fast and the context window small.

  3. Access you can hand to an agent

    TheyAsked has a /mcp/readonly endpoint and 10 granular scopes, so a research agent never even sees write tools. It supports OAuth 2.1 with dynamic client registration and API keys for Claude Code, CI and scripts. Customer text is fenced as untrusted data, destructive tools need an explicit confirm, and every agent action lands in an audit log. Canny's MCP uses OAuth only, does not support dynamic client registration, and gates tools by the connecting user's role.

  4. Prompts and views inside Claude

    TheyAsked ships 5 MCP prompts (weekly_digest, triage_new_requests, prep_roadmap_review, draft_release_notes, account_brief) that show up as slash commands. In claude.ai and Claude Desktop, tools draw interactive MCP Apps views: a ranked ledger, KPI cards, segment breakdowns and request cards with a confirmed "Mark planned". Canny's MCP help article documents neither prompts nor interactive views.

  5. Flat price per workspace, voters never counted

    TheyAsked Pro is $29 a month per workspace with MCP, REST API and webhooks included, and voters are never counted. Canny prices by tracked users (anyone who posts, votes or comments, plus actions by Autopilot), and its MCP needs Pro, which starts at $79 a month billed yearly, with Canny Ideas enabled. A launch that brings in a wave of voters does not change your TheyAsked bill.

  6. Mobile embeds for iOS and Android

    TheyAsked has a web widget plus iOS and Android WebView embeds, with signed JWT SSO so customers vote as themselves without a second login. Votes are tagged by platform, so who_asked_for can tell you whether a request comes from your mobile users or your web users.

TheyAsked vs Canny

Line by lineTheyAskedCanny
Pricing modelFlat per workspace. Voters never counted.By tracked users: anyone who posts, votes or comments, plus admin and Autopilot actions.
Entry price with MCPPro, $29/mo ($24/mo yearly). 14-day Pro trial.Pro, from $79/mo billed yearly, with Canny Ideas enabled.
Free planFree: one board, unlimited posts and voters, roadmap and changelog.Free: 25 tracked users, 5 managers, Autopilot included.
MCP tools33 tools (17 read, 16 write), plus 5 prompts. No tools that delete posts or comments.62 tools (29 read, 14 write, 19 admin), including delete tools.
MCP auth and access controlOAuth 2.1 with DCR, or API keys. 10 granular scopes and a /mcp/readonly endpoint.OAuth only, no dynamic client registration. Tools gated by Canny role; custom-role permissions on Business.
Agent safetyCustomer text fenced as untrusted, confirm on destructive tools, notify off by default, audit log.Role-gated tools. Further safeguards not documented in the MCP help article *
Revenue weightingMRR of voting companies, momentum and opportunity built into the ranking tools.MRR or ARR per idea via Impact Overview, sort by total MRR, opportunity value from Salesforce or HubSpot.
Views inside ClaudeMCP Apps: ranked ledger, KPI cards, segment breakdowns, request and account cards.None documented.
AI feedback captureAgents log requests from calls with create_post and vote_on_behalf; duplicate check with find_similar_posts.Autopilot extracts and dedupes feedback from Gong, Intercom, Zendesk, Zoom, Slack, review sites and more.
MobileWeb widget plus iOS and Android WebView embeds, votes tagged by platform.Widget can be used inside a mobile app. No native SDK documented *
End-user SSOSigned JWT SSO on Pro.Widget SSO token on all plans. Okta, OIDC and OneLogin on Business.
Changelog notificationsPublishing notifies everyone who asked for the shipped requests.Status change emails to voters; changelog emails to subscribers on Pro and Business.

From each vendor’s public pages, checked 2026-10-22. * Not confirmed on Canny’s own site.

When Canny is the better pick

  • Teams whose feedback lives in calls, tickets and reviews: Canny Autopilot extracts and dedupes requests from Gong, Intercom, Zendesk, Zoom, Slack, G2, Capterra, app stores and more, on every plan.
  • A wide integration catalog: Canny connects to Salesforce, HubSpot, Jira, Linear, GitHub, ClickUp, Asana, Azure DevOps, Microsoft Teams, Discord and dozens of feedback sources.
  • Enterprise sales workflows: Canny ties ideas to Salesforce and HubSpot opportunity value and offers Okta, OIDC and OneLogin SSO on Business.
  • A broader MCP write surface: Canny exposes admin tools for changelog entries, segments, views and deletes that TheyAsked deliberately leaves out.

Moving from Canny

Export your posts from Canny as a CSV and drop it into Settings → Import → Canny, or the import step of onboarding. TheyAsked reads Canny's export headers as they are (post title, post detail, category as board, status, tags, score as votes, post author, date created, comment count), and you can remap any column on the preview step. Statuses and boards are mapped on a dry-run preview before anything is written. If your file has a column of voter emails, each email becomes a real vote tied to that person, so revenue and segment ranking work on them; otherwise the vote total is kept as a count. Comments do not come across with the posts CSV. Re-running the import updates what it brought in and never duplicates it.

How importing works

Questions

Is there a free Canny alternative?
Yes. TheyAsked Free costs $0 with one board, unlimited posts and voters, a public roadmap and a changelog. Voters are never counted on any plan. MCP, the REST API and webhooks start on Pro at $29 a month, with a 14-day Pro trial when you sign up.
Does TheyAsked have an MCP server like Canny?
Yes. TheyAsked has 33 MCP tools (17 read, 16 write), 5 prompts and interactive views in claude.ai and Claude Desktop. It connects with OAuth 2.1 (including dynamic client registration) or an API key, and offers a read-only endpoint and 10 granular scopes. It is included from Pro at $29 a month; Canny's MCP needs Pro from $79 a month billed yearly with Canny Ideas enabled.
Can I import from Canny?
Yes. Upload the Canny posts CSV export and TheyAsked brings over posts with their board, status, tags, author, created date and vote totals. A column of voter emails becomes real votes. Comments are not part of the posts CSV, so they do not come across. Imports are previewed first and are safe to re-run.
Can TheyAsked rank feature requests by revenue?
Yes. Pass plan and MRR for your companies through SSO, the API or upsert_company, and every ranking tool weighs requests by the MRR of the companies voting, alongside unique voters and momentum. who_asked_for shows the companies and plans behind one request in a single call.
Does TheyAsked extract feedback from calls and tickets like Canny Autopilot?
Not automatically. TheyAsked has no connector that reads Gong, Zendesk or Intercom on its own. Your agent can log what a customer said with find_similar_posts, then vote_on_behalf or create_post with their words as the note. If automatic extraction from many sources is your main need, Canny is the stronger pick.

Sources

Other comparisons