Turn what you capture into retrievable knowledge

TopoTopic turns recordings, documents, videos, and links into a searchable knowledge base with grounded answers. Capture once; after processing completes, get reports, a 3D knowledge graph, review cards, and voice search. Source links point back to the exact passage, page, or transcript moment they came from so you can verify them quickly.

Start in Memo. After signup, open Memo from the navbar to record live, upload a file, paste from the clipboard, paste a YouTube link, or import a web link. If TopoTopic asks for account details, complete them once and continue to Memo. Topics and keywords stay available later from Topics in your profile menu, and they can help personalize Learn, recommendations, and related workspace surfaces.

Yes — TopoTopic is free during the beta, which is limited to 500 seats. Create a workspace and start capturing in Memo after account setup. When self-service billing is enabled for your account, Settings shows the available Stripe billing controls for starting, changing, or managing a subscription.

Any modern browser, including Chrome, Firefox, Safari, and Edge, on desktop and mobile. Live recording works best on Chrome and Firefox; Safari is supported with gesture-based microphone permission.

Anything knowledge-bearing:

• Live recordings from your microphone, tab, screen, or system audio setup
• Audio and video files
• PDFs, documents, slides, notes, and pasted text
• Images as session photos
• YouTube videos and web links

Supported content you bring in through Memo is transcribed, extracted, or attached to a session, then connected to your workspace knowledge when processing completes.

On the Memo page, choose a live audio source and grant the browser permissions it asks for. TopoTopic can record a microphone, tab or screen audio, tab or screen plus mic, or system audio plus mic when a loopback or monitor device is available. Pause, resume, or add context documents mid-session. When you stop, a final analysis produces summaries, notes, questions, and chapters.

Yes, when Live translation is enabled for your account. Open the Live page from the navbar: the listen-along translator can use your microphone, tab audio, or a local media file, then play translated audio while showing source and translated captions. You can also start a translated call with another person from the same page.

Choose a target language from the currently enabled language list. Availability depends on browser microphone or WebRTC permissions, workspace access, daily live-audio limits, and the server-side live-translation rollout, so the session start is the final availability check.

Yes. Paste a YouTube URL on the Memo page and TopoTopic uses available captions or a transcript you provide. Videos without usable captions may need a supported transcript or upload path. Successful imports become sessions with summaries, keywords, chapters, and searchable transcript text.

Long videos can be imported as clips. Add start and end times in Memo, or paste a YouTube URL that already includes t=, start=, or end=. Each import is limited to 60 minutes, but the resulting session reflects only the selected interval.

Yes. Upload a document or text-oriented file in Memo, wait for the summary to finish, then open the interactive viewer for that upload and ask questions there. The read-only item page is for browsing the record; the interactive viewer is the place to chat with the document.

Document summarization can finish in the background, so it is normal to see the summary appear a little later while you continue using the app.

Transcribed and analyzed the same way as live sessions:

• Audio: AAC, AIFF, MP3, WAV, M4A, OGG, FLAC, WebM, WMA
• Video: MP4, M4V, MOV, AVI, MKV, WMV
• Documents and text: PDF, TXT, Markdown, RTF, DOC, DOCX, ODT, PPT, PPTX
• Images: PNG, JPG, WEBP, HEIC, HEIF

Images become photos attached to a session and should be uploaded separately from documents, audio, or video.

Yes. Use Paste from clipboard in Memo, or paste with the keyboard, to bring in copied links, text, images, or files. Plain text is saved as a text file so it follows the same durable upload and analysis path as other source material.

Clipboard access depends on browser permission. If the browser blocks direct clipboard reading, use keyboard paste, drag and drop, or file browse instead.

Yes. Paste any URL on the Memo page. TopoTopic extracts the page text, or timed transcript where available, creates a session, and runs the full analysis pipeline. This works well for articles, blog posts, and research papers.

Every session can generate four main analysis views, each with a different job:

• Summary: the main takeaway and the core points to remember
• Notes: the concrete details, facts, references, and context worth keeping
• Insights: the non-obvious patterns, implications, and interpretations pulled from the original
• Questions: the follow-ups, unresolved points, and prompts for deeper review
• Chapters with timestamps
• Extracted keywords
• Speaker-aware transcript structure when available

These outputs show up in the session report and can support Vault, Graph, Learn, and related workspace surfaces.

Analysis language is a session setting. By default, Auto-detect writes summaries, notes, insights, questions, chapters, and reports in one detected dominant language. If you choose Keep original languages, TopoTopic keeps the source language or languages, so a mixed-language session may show mixed-language analysis instead of translating everything.

Before recording, adjust Analysis language in Memo Live settings. During review, use the Analysis card, the session analysis settings, or the Rerun analysis dialog to keep original languages or choose a specific output language. Transcripts always stay verbatim.

Yes. From the Library session card, the session preview panel, or the session page itself, use Rerun analysis to regenerate all AI outputs. You can review the analysis language policy before rerunning and separately adjust transcription settings such as vocabulary hints where available.

TopoTopic automatically separates speakers in your recordings. After a session, you can assign real names to each detected speaker. Cross-session speaker recognition only applies if Voice Recognition is enabled in Settings → Privacy; otherwise speaker labels stay session-scoped.

A workspace is TopoTopic's main boundary for captured content, organization, and AI recall. Each workspace has its own sessions, collections, graph, and workspace-level context. Use separate workspaces when you want a clear boundary between personal, team, client, or project material.

Most accounts start with two: Work is your default operating context; Private is a separate personal space. You can create or join additional workspaces from Settings when your role and plan allow.

Most of TopoTopic follows the workspace you currently have selected in the header switcher. Sessions, collections, Library search, Vault answers, graph exploration, Learn material, and AI outputs stay inside that workspace unless a feature explicitly says otherwise. If something seems missing, check the active workspace first.

A few. Direct messages are participant-based, so they can span workspaces. Most other surfaces that expose captured content, people context, collections, or graphs require explicit access to the current workspace and do not silently fall back to another one.

The Library is the central hub for your active workspace. Browse, search, and manage sessions, collections, and content from one place. Use collections to group related sessions, and the search bar to find anything by keyword or meaning.

Collections are folders for your sessions. Create them manually or let TopoTopic auto-cluster related sessions together. Pin your favorites to keep them at the top of the sidebar; pinned collections are also excluded from automatic AI recuration.

Library search combines keyword matching and semantic search inside the active workspace. Type a query and results surface across sessions, collections, documents, notes, tasks, and reminders, ranked by relevance with matching text highlighted.

The Vault answers natural-language questions across the content in your active workspace and returns answers grounded in your own knowledge base, each linked back to its source. Think of it as recall for everything you have captured — scoped to the workspace you are in, so work and private material never mix.

Every answer is tied back to the exact supporting passage in your own material — your notes, sessions, documents, and saved links. TopoTopic shows the page, transcript excerpt, or session reference used to support the answer, so you can inspect the original context instead of trusting a standalone summary.

Library search finds sessions and items by keyword. Vault answers questions: it synthesizes information from multiple sessions and returns an answer with references to the original material.

Yes. TopoTopic has a global voice assistant. Activate it from the microphone icon in the navbar and ask questions naturally. The assistant searches your active workspace in Vault, navigates pages, and reads results aloud.

The Graph is a 3D visualization of every keyword, topic, and connection across your workspace. Navigate it with keyboard, mouse, or voice. Nodes represent keywords; edges show co-occurrence and semantic similarity.

Automatically. Every time you capture a session, the AI extracts keywords and maps them to topics. When keywords appear together across sessions or have similar meanings, the graph connects them. The more you capture, the richer your graph becomes.

Yes. The graph has a built-in search panel. Type a keyword to highlight it and its connections. You can also synthesize explanations of keyword relationships using the AI synthesis feature.

Learn is a personalized review feed made from source-backed notes in your active workspace. It keeps the feed focused on durable notes from sessions and imports, then adapts ordering over time based on what you review, save, dismiss, or revisit.

TopoTopic turns eligible session material into note review cards after quality checks confirm there is enough source-backed substance. The current Learn feed filters out non-note analysis, raw uploads, and low-signal material before cards appear.

Yes. The feed prioritizes content from your selected topics and keywords in Topics from your profile menu. You can refine those preferences after capture, save or dismiss individual cards, and return to the original session when you need more context.

Processed sessions can include a report with versioned summaries, insights, notes, questions, chapters, speaker profiles, and a searchable transcript. Navigate between content versions using left and right arrows to see how the analysis evolved.

Yes, in multiple formats:

• Copy the full report as Markdown
• Export sections as .md, .pdf, or .docx
• Share or copy a session report link when sharing is available

Studio is the planning and review workspace for sessions. Use it before capture to name a session, add keywords, write a briefing, and attach materials; use it after capture to trim excerpts and export edited audio without changing the original transcript or AI analysis.

Use Studio when you want a prepared session before Memo, supporting material attached to the same session, a follow-up planned from session context, or a cleaned-up recording after capture.

No. Studio edits are non-destructive. The original transcript, report, and AI outputs stay intact. Only the exported clip reflects your trims.

Silicon is a workspace dashboard for practical follow-up. It groups notes, tasks, reminders, captures, calendar drafts, and recent activity into clusters so you can see what is on your plate without manually organizing anything.

Use it when you want to turn saved material into concrete next actions, add a quick reminder, or review upcoming work without leaving the workspace.

Silicon groups related items using keyword and embedding similarity. Clusters get descriptive titles derived from the content. When you add or update items, clusters refresh automatically in the background.

Connect is the timeline and people surface for your current workspace context. Use it to post updates, review workspace activity, manage drafts, view your own connections, and open profiles you explicitly search for by handle.

It is designed to keep relationships tied to saved work without turning profiles into a public browsing directory.

Create and filter timeline posts, open related session links when they appear, search people by @handle, and follow up from a profile or connection view.

TopoTopic does not browse or suggest hidden people automatically; sharing and publishing remain explicit actions. Connection and messaging access still depends on your workspace and authentication state.

Inbox supports direct messages, group conversations, and Walter threads. It is also where approval-based assistant actions land, including Google Calendar event drafts that Walter prepares before anything is created.

Yes, when translated calling is enabled for the direct conversation. In Inbox, direct Messages calls can show a Translated call panel with target-language selection, translated playback, original-audio controls, and source and translated captions. Each participant can enable translation for the other participant's audio, so both sides can hear translated playback when both enable it.

The call still uses the normal Messages call flow. Translation availability depends on browser WebRTC permissions, conversation access, enabled target languages, daily live-audio limits, and the same live-translation rollout checks used by the Live page.

The global voice assistant is available from the microphone icon in the top navigation. It can search your Vault, navigate pages, read content aloud, create reminders, and control playback, all by voice.

Yes. On the Topics refinement page, you can say commands like select Machine Learning or add React hooks while choosing topics and keywords.

Yes. It works on any device with a microphone. On Safari, a permission prompt appears on first use.

Primary app storage runs on AWS eu-west-1, Ireland, using encrypted Aurora PostgreSQL and S3. Traffic to TopoTopic uses TLS. Some connected providers and processors may process data in other regions as described in the Privacy Policy and Subprocessors page.

Yes. Go to Settings → Privacy → Deactivate account. The current flow starts a 30-day grace period before deletion. Account-only data is removed from active systems where supported, while shared-workspace records, security/legal records, backups, and copies already exported or shared may remain as described in the Privacy Policy.

Google sign-in and passkeys are the fastest ways to get started. Email and password is also supported. If additional sign-in methods are available for your account, they will appear on the sign-in screen.

Check that your browser has microphone permission for topotopic.com. On Safari, the permission prompt only appears after a user gesture, so click the record button first. If you use a Bluetooth device, make sure it is connected before starting.

Processing time depends on source length, file type, provider availability, and queue load. If analysis appears stuck, try Rerun analysis from the session card in Library; larger imports may take longer, and the session page updates when the server finishes.

Check that you are in the correct workspace in the header switcher. Sessions are workspace-scoped, so a session recorded in one workspace will not appear in another. You can also use Library search to find sessions by keyword or content inside the active workspace.

The graph builds from your captured content. Capture a few sessions or import some documents and keywords and connections will start appearing. The more diverse your content, the richer the graph.

Check that your browser is not blocking downloads for topotopic.com and retry from the session page in Library. Large sessions may take a minute to package before the download starts.

Confirm the browser has microphone permission, then refresh the tab. Modern browsers such as Chrome, Firefox, Safari, and Edge are supported; on Safari, the first use requires a click before the mic activates.

Long files and queued imports can take longer to finish. Keep the session page open or return later; completed analysis appears automatically when ready.

Verify your card in Settings → Billing and retry. If the issue persists, contact support.