Feature Ideas

Most of my meetings are virtual and I'm wearing a headset. This makes pocket way less useful for me as it's unable to record meetings that are not happening live or if it cannot hear the other people speak. A pocket app that can listen in on meetings or be able to import transcripts would be awesome

I’ve been heavily using Pocket Record Conversations inside operational workflows, meetings, strategy sessions, assessments, and execution tracking, and honestly it’s become one of my favorite tools for capturing organizational memory and conversations in real time. There are a few friction points I’ve run into around speaker identification and transcript organization, but they honestly feel less like complaints and more like opportunities for future evolution because the platform already has so much potential. One of the biggest workflow challenges right now is speaker attribution in larger meetings and recurring team conversations. The biggest underlying issue is that multiple real people can accidentally get grouped under a single speaker identity, which makes transcript cleanup extremely manual and difficult at scale. Right now, the workflow is mostly: • Rename one speaker globally or • Manually correct transcript sections one at a time The problem is that “rename all” can unintentionally relabel multiple actual people because they were initially merged together under the same speaker label. Once that happens, there’s currently no efficient way to split those speakers back apart across the transcript on either mobile or desktop. This becomes especially difficult in long operational meetings where speaker accuracy matters and recurring voices should ideally become more identifiable over time. A few things that would make a huge difference: • Better recurring voice recognition and speaker fingerprinting • True post-transcription speaker splitting • Ability to isolate and reassign selected transcript ranges without affecting unrelated sections • Better speaker reassignment workflows across both desktop and mobile • Speaker learning/profiles for recurring teams • Confidence indicators where attribution may be uncertain • Dedicated transcript review/validation mode Another area that feels like a major opportunity for growth is the organizational and file management side of the platform. As conversation volume scales, organization becomes just as important as transcription itself. I’d love to see deeper file and folder organization capabilities evolve, especially for users managing large numbers of operational conversations, recurring meetings, assessments, and cross functional workflows. Things that would make a huge difference there: • Nested folder structures/deeper file hierarchy • Bulk tagging and categorization tools • Multi-select organization workflows • Easier grouping of related conversations and recurring meeting threads • Better filtering/search across tags, speakers, projects, or teams • Ability to organize identified speakers or recurring groups more efficiently • Faster bulk management for large transcript libraries The platform already has really strong foundational functionality. It mostly feels like the next evolution is making large scale workflow management, transcript organization, operational retrieval, and institutional memory more scalable and intelligent. Really impressive product overall and excited to see where the platform evolves next. Curious what features or workflow improvements other people would love to see too.

It’s possible to bump the record button or even click to record and then change your mind and cancel before any content is recorded. The system still processes and saves this. It would be good to have a setting that automatically deletes this if no content was recorded or a recording is less than “0” seconds long. For future hardware, it might be good to have a toggle lock to avoided accidental button presses.

Timeout for the LED status light after say 10 minutes of no use. It can draw unnecessary attention and can be a disturbance at night even though it isn't super bright.

Unintentionally recorded conversations when device is in my pocket. Would like the option to 'lock' it to avoid that.

Instead of updating each speaker manually, Pocket should recognize speakers with that voice and offer to update all speaker blocks for that speaker instead of one by one. A one-hour meeting can have 30+ speaking blocks from each party, making it impossible to manually go through the entire transcript to update them all.

I'm sure that a high proportion of your users have iPhones. An integration with Apple apps like Contacts,Calendar, Mail and Reminder would be very productive. I know this isn't easy because of Apple's security restrictions, however I dont trust many web apps because of their insecurity.

Pocket transcripts are often inaccurate or off-tone, which makes sharing them with clients risky and erodes trust instead of reinforcing authority. TL;DR The current workaround is manual: copy your lines from Pocket, paste into GPT, refine for clarity/professionalism, then paste back, which adds time and friction and does not scale. The proposed solution is a “Refine Transcript for Clarity” feature with: a dual-layer transcript (original + refined), a one-click refine button that improves clarity/tone/grammar while preserving meaning, and an edit-and-approve flow. After refinement, Pocket should let users regenerate summaries, notes, and other outputs from the cleaned transcript so shared links are client-ready and trustworthy. This would increase user trust, professional image, shareability, retention, and create a viral loop, especially for sales, consulting, and client meetings, where it acts as a client-facing trust layer. “Dual-Layer Transcript: Original + Refined Version for Client-Facing Links” Pocket is powerful… but the moment you share a link with a client or prospect and they click into the transcript, things can get uncomfortable fast. Transcripts are often: slightly offmisrepresenting what was actually saidlacking clarity in tone and intentnot aligned with how we actually communicate professionally In a sales or client-facing scenario, this matters. Because now instead of reinforcing authority… it can unintentionally create friction or doubt. What I'm Already Doing (Manual Workaround) Right now, the workaround looks like this: Open Pocket on webCopy your speaking linesPaste into GPTRefine for clarity and professionalismPaste back into Pocket We’re literally fixing transcripts manually before sharing them. That’s time, friction, and missed scalability. 💡 The Opportunity (This Is a No-Brainer Feature) Instead of trying to make transcription perfect… give users control to refine what was captured. Proposed Feature: “Refine Transcript for Clarity” Add a simple, powerful layer: Dual-Layer Transcript System Original transcript (unchanged, for reference)Refined transcript (clean, readable, client-ready) One-Click “Refine” Button Refines selected lines or full transcript Improves:claritytonegrammarprofessionalismPreserves original meaning and intent Edit + Approve Flow Users can tweak refined text if needed Gives control without losing speed 🔥 Regenerate Outputs From Refined Version This is the multiplier. After refinement, allow users to: regenerate summariesregenerate meeting notesregenerate frameworks / outputs Based on the cleaned transcript, not the flawed one Why This Will Blow Up (For Users + Pocket) This hits multiple high-impact levers: ✔ Trust Users feel confident sharing links again ✔ Professional Image No more “that’s not what I meant” moments ✔ Shareability More links sent = more exposure for Pocket ✔ Retention Once users rely on refined outputs, switching cost increases ✔ Viral Loop Better shared experiences → more inbound curiosity → more users Real Talk Transcription will never be 100% perfect: background noiseinterruptionsoverlapping voices That’s normal. But refinement bridges the gap between raw capture and real communication. 🚀 Why Founding Members Should Care If you’re using Pocket for: sales callsconsultingclient meetingsdeal flow Then this is not a “nice to have”… This is a client-facing trust layer. Call to Action If you’ve ever: hesitated to send a transcriptedited one manually before sharingor felt like it didn’t fully represent you 👉 Upvote this. Let’s help Pocket evolve from powerful tool → polished, client-ready platform.

Based on this post :https://community.heypocket.com/t/i-created-a-new-calendar-item-it-duplicated-it-and-added-to-unknown-calendar-video-attached/525/4?u=sm7 I'd like to be able to select which calendar to add an event to. I currently have 3 calendars Pocket can read from. If I create a new event, I want to select which calendar to the event to

I personally cannot stand emailed OTP codes. Traditional sign in options like password, passkey and 2fA app codes all offer a superior user experience. I hope to see these sign in methods added!

The device is listed as HIPAA compliant, but it doesn't appear to be so as you can fairly easily download audio recordings off the device simply by connecting it to a computer via USB. This needs to have a password of some sort in order to access.

adding picture support to be added to a summary just highlights and strengthens a conversation.

For an added layer of security it would be nice to have a folder system that can be locked down for privacy. Secured by device passcode or 2FA options.

im very irritated that just by pickimg up and using my phone for other things, I inadvertently activate and then I have several recordings that are 2,3 or 4 hours long so I would like a way when it is attached to the phone to turn it on only by the app. And maybe a quick and easy way to select what recordings to sync because I don’t want to wait for what I know is garbage to have to sync to the phone.