Private previewWorkspaces are by invitation

HeedPost

01Collect

Feedback boardsBoards that ask for exactly what your team needs to act.Votes, comments & mentionsOne vote per person, threads with mentions, and notes only your team sees.In-app widgetFeedback from inside your product, from customers who are already signed in.

02Understand

Merge & duplicatesTen posts asking for the same thing become one, with every vote kept.User segments & revenueProSee which requests your biggest customers want, in dollars.InsightsWho reads, what's rising, and whether you answered.HeedPost AIPlus and ProTags, duplicate judging, drafts, summaries, themes and Ask, each with its own switch.

03Ship and tell

Roadmap & statusesMove a post and everyone who asked hears, in your own status words.ChangelogRelease notes linked to the requests behind them, mailed to everyone who asked.Notifications & digestThe right email to the right person, and a switch for every one.

04Run it

Activity logEvery change, who made it and when, a day at a time.Team, roles & sign-inUnlimited teammates, clear roles, and sign-in your way.Branding, domain & languagesYour logo and colour, your own domain, and 21 languages.ImportMove in from Canny, Featurebase or UserJot, with nobody emailed.

05Connect

Slack & issue trackersPlus and ProSlack threads, and issues in Linear, Jira, GitHub or ClickUp that move posts back.MCP serverPlus and ProYour AI agent can read, file, vote and ship, as you.REST APIPlus and Pro27 operations and one key: build anything on top of your feedback.WebhooksPlus and Pro110 events, signed and retried, the moment they happen.
AIDocsHelpPricingBlog

Collect

Understand

Ship and tell

Run it

Connect

Request an inviteSign in

Notes/Building on it

Letting an AI agent work your feedback board

Agents can file, triage and ship from the tools your team already uses. A few principles let them in safely.

Nº 35 · HeedPost team · Oct 18, 2026 · 2 min read

HEEDPOST NOTESNº 35

AI assistants can now do more than answer questions. Connected to your tools, they can act: look something up, file a ticket, change a status. For a feedback board, that means someone can ask "which ideas under review have the most votes?" from the assistant they already use, and then say "move the top one to planned and tell whoever asked".

That's genuinely useful. It also means giving software the ability to change things your customers see. A few principles make it safe.

Give it a person's permissions, not more

An agent should act as a specific person on your team, with exactly that person's permissions. If the person can't delete a board, neither can the agent. If they're a member rather than an admin, so is the agent. That way there's no separate set of agent rules to reason about. You already know what each person can do.

Start with reading

The first useful thing most teams do with an agent is ask questions. What's rising this month? What did this customer ask for? What shipped last quarter? None of that changes anything. Start there, and ask it to change things once you've seen how your team uses it.

Expect it to be wrong sometimes

Agents misread requests and pick the wrong board or status. Good tools make that recoverable. When an agent asks for something that doesn't exist, it should get a clear sentence back, like "there's no status called Shiped", so it can correct itself instead of failing silently.

Keep a record of what it did

Every change an agent makes should be visible, with whose key made it and when. If an agent moves thirty posts by mistake, you want to see exactly which thirty, and undo them.

Keep the human moments human

Some things should stay with a person even when an agent could do them. Replying to an upset customer. Closing a request with a reason. Announcing a release that a hundred people waited for. An agent can draft those. A person should send them.

One door, the same rules

The simplest way to keep agents safe is to give them the same door as everything else. In HeedPost, an agent's tools are the same operations as the public API, with the same limits and the same checks. There's no special path that skips the rules. If it's safe through the API, it's safe through an agent, and if something isn't allowed for one, it isn't allowed for the other.

Start hearing everyone.

Free for your first two boards, with no card.