This Privacy Policy explains how Todoki ("Todoki," "we," "us," or "our") handles personal information when you use our websites, applications, meeting intelligence services, and related features (collectively, the "Service"). By using the Service, you acknowledge the practices described below.
1. Information we collect
Account information
When you create an account or sign in with Google, we may receive your name, email address, profile image, account identifier, and authentication information. We store session and workspace membership information needed to operate your account.
Meeting and workspace content
We process information you or your workspace submit, including meeting links, titles, participants, transcripts, recordings or recording references, summaries, decisions, action items, comments, bookmarks, destinations, and execution results. Meeting content may contain personal information about participants and others.
AI processing
When AI extraction or meeting answers are enabled, we send relevant transcript turns or matching excerpts to OpenRouter and its selected model provider to generate summaries, outcomes, or answers. With your live copilot processing consent, requesting a recap or asking a question sends the question and a bounded recent transcript window to those providers. Live answers are drafts displayed privately in Todoki; they do not automatically speak, dispatch tasks, or become Walrus memory. Requests are processed under those providers' terms and privacy practices. Todoki does not save Ask Meetings or live copilot questions and answers as separate database records.
Private Vault
Private Vault is a separate browser-encrypted context mode. Your title, filenames and sources are encrypted locally using AES-256-GCM with a passphrase-derived key before upload. The passphrase and derived key are not transmitted to Todoki, Walrus or cloud AI. Todoki temporarily queues ciphertext, forwards ciphertext through the managed Walrus relayer, and clears the queued ciphertext from the active database row after hash-verified storage and retrieval. We retain ownership, byte size, hash, storage pointers, status and payment metadata; backups may retain ciphertext. The relayer may embed and add storage encryption to ciphertext without seeing your underlying sources.
Vault decryption and source search run in your browser. Cloud AI is optional and requires explicit approval for each request: your question, reviewed plaintext excerpts and any recent dialogue you choose to include are transmitted through Todoki to OpenRouter and its selected model provider. Those services can read the shared text for processing, and their data policies apply. Original filenames, the full vault, the passphrase and the derived key are not included by this feature. Todoki does not save these questions, excerpts or drafts as separate database records. Cancelling or locking cannot retract data already processed by providers. We cannot recover a lost vault passphrase. Keep a strong unique passphrase and an encrypted backup. Walrus stores publicly retrievable ciphertext; removing app references does not erase every distributed copy. An unlocked vault is accessible to code on your device, so malicious extensions, compromised application code or an infected device can expose it. Browser/OS memory erasure is not guaranteed. These protections do not change existing cloud context, meeting capture, transcript, or managed outcome-snapshot processing.
Connected services
If you connect Google, Slack, or Linear, we receive authorization tokens, account identifiers, scopes, and metadata needed to provide the integration. When you approve an action, we send relevant content to the selected service, such as a Slack message, Gmail draft or message, or Linear issue. Google API data is used only to provide and improve user-facing integration features and is handled in accordance with the Google API Services User Data Policy, including its Limited Use requirements.
Google Calendar auto-join
When you choose to connect Google Calendar, Todoki requests read-only calendar event access and offline access so the connection can refresh while you are away. We currently read your primary calendar to identify upcoming meetings, including event identifiers, titles, start and end times, timezone, meeting links, and whether you host or have accepted the event. We store the event metadata needed for scheduling and show upcoming meetings to members of your Todoki workspace. Private event titles and links are masked. We use calendar change notifications to update schedules and do not create, edit, or delete Google Calendar events.
Connecting a calendar does not enable recording. You choose an auto-join rule and confirm recording permission before Todoki schedules a bot. For the meetings you enable, we send the meeting link, scheduled time, bot name, and an internal session identifier to Recall.ai to join and capture the meeting. You can skip meetings, pause auto-join, or request disconnection. Disconnection stops calendar bots and removes the saved calendar tokens after outstanding bot requests are resolved; already captured meeting content and stored event records may remain. You can separately revoke Todoki in your Google Account permissions and request deletion using the contact details below.
Meeting capture
If you dispatch a meeting notetaker, we provide Recall.ai with information needed to join and process the meeting, such as its URL, bot name, schedule, and session identifier. Recall.ai may process audio, video, participant, and transcript data on our behalf.
Optional Walrus Memory
When enabled by the operator, a meeting owner may explicitly save a versioned outcome snapshot to Walrus Memory. It includes the meeting title, date, summary, reviewed actions, confirmed approved decisions, and their supporting quotes. Full transcripts and unreviewed actions are excluded. The hosted Walrus Memory relayer receives plaintext for embedding and encryption; stored blobs use Seal encryption. The initial integration uses Todoki's operator-owned memory account and delegate, and access through Todoki follows the meeting's current sharing permissions. The account operator and relayer are part of this processing trust boundary. Stopping Todoki recall or deleting a local meeting removes access through Todoki but does not erase distributed blobs, existing copies, or provider backups. Memory search questions are sent to the relayer when you use that feature.
Wallet payments
If you choose a wallet payment, we process the public payer and receiving addresses, network, amount, transaction digest, and payment status to fulfill and reconcile your request. On-chain transfers and associated public identifiers can be visible permanently on the selected blockchain. Your wallet authorizes each payment; Todoki does not request your wallet seed phrase or private key.
Technical information
We and our infrastructure providers may collect IP address, browser and device information, request and diagnostic logs, timestamps, cookie data, and Service usage. We use cookies and similar technologies necessary for authentication, security, and core functionality.
We use Vercel Web Analytics to understand page visits and usage patterns. We remove query strings and URL fragments from page-view URLs before sending them, and do not send meeting transcripts or task content as analytics events.
2. How we use information
We use personal information to:
- Provide, maintain, secure, and troubleshoot the Service.
- Authenticate users and manage accounts and workspaces.
- Capture and organize meetings, produce transcripts and insights, and support search.
- Connect integrations and execute actions that users review and approve.
- Communicate, respond to requests, detect abuse, comply with law, and improve performance.
We do not sell personal information. We do not use Google user data for advertising.
3. How we share information
We may share information:
- With service providers for hosting, databases, authentication, monitoring, infrastructure, AI processing, and meeting capture, including Vercel, Google Cloud, managed PostgreSQL providers, OpenRouter and its selected model provider, and Recall.ai.
- With connected services such as Google, Slack, or Linear when you connect them or approve an action.
- Within your workspace according to member permissions.
- For legal and safety reasons to comply with law, protect rights and safety, investigate abuse, or enforce agreements.
- In a business transaction such as a merger, financing, acquisition, reorganization, or asset sale, subject to appropriate safeguards.
4. Retention and deletion
We retain information as needed to provide the Service, maintain legitimate business records, resolve disputes, enforce agreements, and meet legal obligations. Retention varies by data type and workspace configuration. Disconnecting an integration stops future access but may not delete content already created in Todoki or the connected service.
You may request access, correction, export, or deletion by emailing hello@todoki.org. We may need to verify your identity. Workspace administrators may also control workspace content. Some information may remain in backups or be retained where required by law.
5. Security
We use reasonable administrative, technical, and organizational safeguards designed to protect information. Integration credentials are encrypted at rest by the application, and production traffic should use HTTPS. No transmission or storage method is completely secure, so we cannot guarantee absolute security.
6. Your responsibilities and choices
You are responsible for providing required notices and obtaining all permissions or consents necessary before recording, transcribing, or uploading a meeting. Recording and consent laws differ by location. You can choose not to connect integrations, disconnect them from Todoki, and revoke access through the provider.
7. International processing and children
Information may be processed in countries other than where you live, including the United States, where privacy laws may differ. The Service is intended for business users and is not directed to children under 13. Contact us if you believe a child has provided information to the Service.
8. Changes and contact
We may update this Policy as the Service or law changes. We will post the revised version here and update the effective date. Questions or privacy requests may be sent to hello@todoki.org. For service questions, contact hello@todoki.org.