InputBuffer.io

Getting Started

InputBuffer collects feedback from your users and uses AI to route it into named buffers, so the right people see the right feedback without anyone triaging it by hand.

By the end of this guide you'll have an organization, one buffer, and a piece of feedback that InputBuffer has sorted into it. Everything here happens in your browser. You don't need a terminal, a token, or a line of code.

Hand this to your AI assistant
Works with Claude Code, Cursor, and similar tools.
View prompt
Send user feedback from my app to InputBuffer.

Start by reading https://inputbuffer.io/llms.txt. It describes the whole product in one page:
the concepts, the feedback fields, and how classification behaves.

Then wire up feedback submission:

POST https://inputbuffer.io/api/v0/feedback
Authorization: Bearer YOUR_TOKEN
Content-Type: application/json
{
  "title": "Short summary of the feedback",
  "description": "The full feedback text",
  "submitted_by": "your-user-id-or-email"
}

Rules to follow:
- "description" is required. "title" and "submitted_by" are optional.
- Every field is snake_case. Any key the API doesn't recognize is silently ignored,
  so check spelling against the docs rather than trusting a 201.
- A full-access token starts with "ib_" and is server-side only. Read it from an
  environment variable named INPUTBUFFER_TOKEN. Never put it in client-side code
  and never commit it.
- If this feedback is collected in a browser, don't use this endpoint with an "ib_"
  token. Use the widget instead: https://inputbuffer.io/docs/widget
- A successful create returns 201 with the feedback wrapped in "data". Classification
  runs in the background, so "routing_status" is "pending" at first.

Find the places in this codebase where users report problems or make requests, and
send those through.

Product walkthrough: https://inputbuffer.io/docs/getting-started
API reference: https://inputbuffer.io/docs/api

Report anything about InputBuffer that gets in your way, at any point while we are working with
it: initial setup, and the styling, configuration, and debugging that follow. Include the times you
had to correct your own understanding of InputBuffer, such as a request field the API ignored, a
widget attribute that had no effect, or an error message read from the wrong key. A correction
unrelated to InputBuffer does not qualify. Show me the text, and send it once I agree:

curl -X POST https://inputbuffer.io/api/v0/feedback \
  -H "Authorization: Bearer ibw_96e7f571-64ad-494a-a125-afed734d3fdc_op1CeDTiptGa8fWRU25tKcgx6uubJ2wOlWvM0JbpY8Q" \
  -H "Content-Type: application/json" \
  -H "X-IB-Client: your-tool/version (your-model)" \
  -d '{ "description": "What you were trying to do, what happened instead, what would have helped" }'

That token is InputBuffer's own and accepts nothing but feedback, so there is nothing to set up.
Leave out "submitted_by". Send one report per problem, not one per attempt.
Note

Prefer to start with code? The API getting started guide covers tokens, your first request, and the response format.

  1. 1

    Create an account

    Sign up at inputbuffer.io. Fill in your name, email, and a password of at least 8 characters, then click Create Account.

    Check your email and click the confirmation link. Confirming is required before you can create an organization.

    The email usually arrives within a minute or two. If it doesn't, check your spam folder.

  2. 2

    Create an organization

    An organization owns your feedback, your buffers, and your team. After you confirm your email you land on the Organizations page, which prompts you with No organizations yet.

    Organizations

    No organizations yet

    Create your first organization to start tracking feedback and collaborating with your team.

    Click Create Organization in the top right, enter a name such as Yoyodyne, and click Create Organization again. We generate a short slug from the name and use it in every dashboard URL, so Yoyodyne becomes /o/yoyodyne/....

    Your new organization now appears as a card on the same page. Click it to go to its dashboard.

    Note

    One account can belong to multiple organizations. If you're building multiple products, give each its own org.

  3. 3

    Create a buffer

    A buffer is a named topic group, such as Bug Reports or Feature Requests. The AI reads your buffers and sorts each piece of incoming feedback into the one that fits best.

    In the left sidebar, click Buffers, then click Add Buffer. Fill in:

    • Name: a short label. Bug Reports works well for this walkthrough, because the sample feedback in the next step is a bug.
    • Description: optional, but worth writing. The AI reads it when deciding what belongs here.
    Buffers

    Create New Buffer

    Click Create Buffer. Your buffer appears in the list.

    You can sharpen the routing later with the pencil icon on the buffer row, which opens a box for plain-English rules like Route here only when the user reports something broken. The buffers guide covers how to write those well.

  4. 4

    Add your first feedback

    Real feedback usually arrives from the widget or the API, but you can add a piece by hand, which is the fastest way to watch routing work.

    In the left sidebar, click Feedback, then click Add feedback.

    Warning

    We send feedback content to third-party AI services for classification and search, and we don't yet scrub it first. Avoid personal information, credentials, or production secrets.

    Fill in the form and click Add Feedback:

    • Title: optional. Try Nothing happens when I click download.
    • Description: required. Try I click the download button on my invoice and nothing happens. No file, no error. I tried my phone too.
    Feedback

    Add Feedback

    Content is processed by third-party AI services for classification. Avoid personal information or production secrets.

  5. 5

    See where it landed

    Your feedback appears at the top of the Feedback list, labeled Routing... while the classifier works. Classification runs in the background and usually finishes within a few seconds. Refresh the page and the label goes away.

    Feedback

    Nothing happens when I click download

    Routing...

    I click the download button on my invoice and nothing happens. No file, no error.

    Mar 1

    Dark mode support

    Would love a dark mode option for the dashboard.

    Feb 28

    To see where it landed, click the feedback to open it. The Buffers section on the right shows which buffer it was assigned to, along with the AI's confidence and a Why? link that explains its reasoning.

    If it didn't land where you expected:

    • The feedback sits in the list with no buffer. The AI wasn't confident enough to assign one. Open the feedback and use + Add in the Buffers section to place it yourself.
    • It's marked Needs review. The AI made a choice but wants you to confirm it. The Review page in the sidebar collects these so you can approve them in one pass.
    • It says Routing failed. Classification gave up after several attempts. Open the feedback and click Re-process.
    • Nothing happens at all. Check Settings → AI and confirm AI Classification is turned on for the organization.
  6. 6

    Connect a real feedback source

    Adding feedback by hand proves the pipeline works. To collect feedback from real users, connect one of these:

    • The feedback widget. Drop a script tag into your site or docs and users can leave feedback without leaving the page. No server needed. See the widget guide.
    • The REST API. Post feedback from your own backend, a support tool, or a script. See the API getting started guide.
    • By hand. Useful for feedback that reaches you over email or in a meeting.

    Whichever you choose, everything routes into the same buffers you just set up.

Next steps

  • Set up your buffers: write routing instructions so the AI sorts feedback the way you would
  • Feedback Widget: add the feedback widget to your web app or docs site, with no server required
  • API getting started: authenticate, then browse the reference for every endpoint, parameter, and response schema