NeuroStack

v0.11.1
  • Features
  • Science
  • CLI
  • MCP
  • Install
  • Cloud
  • Connect
  • Compare
  • Blog
Stars38

Privacy Policy

Effective date: 25 March 2026
Last updated: 25 March 2026
Contact: hello@neurostack.sh
Website: https://neurostack.sh

NeuroStack is an open-source (Apache-2.0) knowledge management tool created by Raphael Southall. This policy explains what data NeuroStack collects, how it is processed, and your rights regarding that data.

1. Overview

NeuroStack operates in three modes, each with different data implications:

ModeData leaves your machine?Account required?
Local (Lite/Full/Community)NoNo
CloudYes, when you explicitly pushYes
Remote MCPSearch queries onlyYes

The core principle: in local mode, NeuroStack collects zero data. No telemetry, no analytics, no crash reporting, no phone-home behaviour of any kind.

2. Local Mode (Lite / Full / Community)

What happens

  • Your Markdown vault files are read and indexed into a local SQLite database.
  • Vault files are read-only -- NeuroStack never modifies your source files.
  • The database is stored at ~/.local/share/neurostack/.
  • In Full mode, vault text is sent to a local Ollama instance running on localhost for embeddings and summaries. This traffic never leaves your machine.

What is collected

Nothing. There is:

  • No telemetry
  • No analytics
  • No usage tracking
  • No crash reports sent externally
  • No network calls to any external service

How to delete your data

Delete the database directory at ~/.local/share/neurostack/. Uninstalling NeuroStack removes all traces.

3. Cloud Mode

Cloud mode is entirely opt-in. You must explicitly run neurostack cloud push to upload any data.

3.1 What data is uploaded

When you run neurostack cloud push, changed .md files from your vault are uploaded via HTTPS to a processing server.

3.2 How data is processed

  • Compute: Google Cloud Run (us-central1) receives and processes your files.
  • AI processing: Google Gemini API processes your vault text to generate embeddings (gemini-embedding-001) and summaries/triples (gemini-2.5-flash).
  • Storage: The resulting indexed SQLite database is stored in Google Cloud Storage (gs://neurostack-prod, southamerica-east1), isolated under your user prefix (vaults/{user_id}/).

3.3 Data retention

  • Uploaded vault files are not retained after indexing completes. Only the processed SQLite database is stored.
  • Your indexed database remains in cloud storage until you delete your account.
  • You can download your database at any time with neurostack cloud pull. After download, all searches run locally.

3.4 Server-side queries

When you use neurostack cloud query, your search query is sent to Cloud Run, which searches against your cached indexed database and returns results. No vault files are transmitted during queries.

4. Remote MCP Mode

When you connect an AI assistant (Claude, ChatGPT, etc.) to https://mcp.neurostack.sh/mcp:

  • Authentication is handled via OAuth 2.1 wrapping Firebase Auth (Google sign-in).
  • Search queries are sent to Cloud Run and executed against your indexed database.
  • No vault files are transmitted during MCP queries -- only search queries and search results.

5. Authentication and Accounts

When you create a NeuroStack account (required for Cloud and Remote MCP modes only):

Data collectedPurposeWhere stored
Email addressAccount identification, communicationFirebase Auth, Firestore
Display nameAccount displayFirebase Auth, Firestore
User IDInternal identifierFirebase Auth, Firestore
Tier levelBilling and feature accessFirestore
Account creation dateRecord keepingFirestore
API key hashes (SHA-256)API authenticationFirestore
Usage countsEnforcing tier limitsFirestore

We never store plaintext API keys. Only SHA-256 hashes are retained. Authentication uses the device code flow for CLI login.

6. Billing

Billing is handled entirely by Stripe (operating under SolidPlus LTD).

  • NeuroStack never sees, stores, or processes your payment card numbers or bank details.
  • Stripe processes payments in accordance with Stripe's privacy policy.
  • NeuroStack stores only your subscription tier level and usage counts.

7. Third-Party Sub-Processors

The following third parties process data only in Cloud and Remote MCP modes:

Sub-processorPurposeData processedLocation
Google Cloud RunCompute / indexing / queriesVault files (during indexing), search queriesus-central1
Google Cloud StorageDatabase storageIndexed SQLite databasesouthamerica-east1
Google Gemini APIEmbedding and summary generationVault text (during indexing)Google API infrastructure
Firebase AuthAuthenticationEmail, display name, OAuth tokensGoogle infrastructure
Cloud FirestoreUser recordsUser profile, API key hashes, usage dataGoogle infrastructure
StripePayment processingPayment information (handled by Stripe directly)Stripe infrastructure

Google's data processing terms apply to all Google Cloud services listed above. See Google Cloud Data Processing Terms.

8. Data Security

  • All data in transit is encrypted via HTTPS/TLS.
  • Cloud-stored databases are tenant-isolated by user ID prefix.
  • Database downloads use signed URLs with limited validity.
  • API key authentication uses constant-time comparison to prevent timing attacks.
  • Vault files are not retained after indexing -- only the derived database is stored.

9. Your Rights

You have the right to:

  • Access your data -- download your indexed database with neurostack cloud pull.
  • Delete your data -- run neurostack cloud delete-account to permanently delete your account, all stored databases, and all associated records.
  • Export your data -- your vault files remain on your local machine at all times. The indexed database can be downloaded at any time.
  • Withdraw consent -- stop using cloud features at any time. Your local installation continues to work independently.

For users in the EU/EEA, you additionally have the right to:

  • Rectification -- request correction of inaccurate personal data.
  • Restriction -- request restricted processing of your data.
  • Portability -- receive your personal data in a structured, machine-readable format.
  • Object -- object to processing of your personal data.
  • Lodge a complaint with your local data protection authority.

To exercise any of these rights, contact hello@neurostack.sh.

10. International Data Transfers

Cloud mode processes data on Google Cloud infrastructure in the United States (us-central1) and stores indexed databases in South America (southamerica-east1). For users in the EU/EEA, transfers to the US are covered by Google's Data Processing Addendum, which includes EU Standard Contractual Clauses. See Google's compliance documentation.

11. Cookies

The NeuroStack dashboard at app.neurostack.sh uses only essential cookies for Firebase Auth session management. There are:

  • No tracking cookies
  • No third-party advertising cookies
  • No analytics cookies

12. Children's Privacy

NeuroStack is not intended for use by anyone under the age of 13. We do not knowingly collect personal information from children under 13. If you believe a child under 13 has provided personal information, contact hello@neurostack.sh and we will delete it.

13. Changes to This Policy

We will update this policy as NeuroStack evolves. Material changes will be announced via the project's GitHub repository and changelog. Continued use of cloud features after changes constitutes acceptance of the updated policy.

14. Contact

For privacy questions, data requests, or concerns:

  • Email: hello@neurostack.sh
  • GitHub: https://github.com/raphasouthall/neurostack
NeuroStackApache-2.0 · Built by Raphael Southall
hello@neurostack.shInstallDocsBlogPrivacyDPA