How does MnemonX protect my data?

MnemonX encrypts your memory wallet with AES-256-GCM on your device before it is sent to the server, and the server stores only ciphertext at rest. Your key is also kept server-side so your wallet can sync across your devices and serve requests you or your connected apps make; decryption happens only on those authorized paths, never as ad hoc or support-initiated access.

Which AI tools does MnemonX work with?

MnemonX works with any AI platform that supports OAuth 2.0 integrations, including ChatGPT, Claude, Gemini, Perplexity, and any custom AI tools via the MnemonX Memory API.

Can I control which AI apps access which data?

Yes. MnemonX supports domain-scoped OAuth permissions — you can grant a specific AI app access to only your Work domain, for example, while keeping your Personal domain private.

Is my data synced across devices?

Yes, on paid plans your encrypted wallet is synced to the cloud so you can access it from any device. The wallet is encrypted before it is stored. Your key is also kept on our servers, locked with a separate master key kept outside the database, so it can sync to your devices and serve the requests you make. This is not end-to-end or zero-knowledge encryption.

MnemonX — Your AI Memory, Everywhere

1. Overview

This SLA describes the availability, support, and performance commitments that come with the Enterprise plan, across all live data regions. Free, Pro, Premium and Team plans do not include an SLA: we run them with the same care and publish live status at mnemonx.ai/status, but there is no guaranteed uptime, response time or service credit. They can still contact [email protected].

2. Uptime Commitment (Enterprise)

Plan	Monthly Uptime	Max Downtime/Month	Applies to
Enterprise (Custom)	99.95%	~21 minutes	All regions incl. pipeline

Uptime measured as the percentage of minutes in a calendar month during which the MnemonX platform is available and responding to authenticated API requests. Measured per region independently.

3. Regional Uptime — US vs EU

🇺🇸 United States (us-east-1)

Supabase + Vercel on AWS us-east-1

Maintenance window: Sundays 02:00–04:00 EST

Incident response: English · US Eastern hours primary

🇪🇺 European Union (eu-central-1)

Supabase + Vercel on AWS eu-central-1 (Frankfurt)

Maintenance window: Sundays 02:00–04:00 CET

Incident response: English · CET business hours primary

Planned maintenance: We provide at least 48 hours notice for planned maintenance via the status page and, for enterprise customers, direct email notification. Planned maintenance during the maintenance windows above does not count against uptime calculations.

4. Support Response Times (Enterprise)

Plan	Response Time	Channel	Hours
Enterprise	2 hours (P1) / 4 hours (P2)	Dedicated Slack + Email	24×7 for P1

Other plans get support by email on a best-effort basis (Pro: standard queue; Premium and Team: priority queue), with no guaranteed response time. EU enterprise customers: first-response engineers are available during CET business hours (09:00–18:00 CET, Mon–Fri) for non-P1 issues.

5. Data Residency SLA

5.1 EU data stays in EU

MnemonX guarantees that EU-region encrypted wallet content and encryption keys will not be transferred outside EU/EEA infrastructure without explicit consent or legal requirement. This is a contractual commitment backed by our DPA and SCCs.

5.2 Regional fault isolation

US and EU regions operate on independent Supabase projects. An outage in the US region does not affect the EU region and vice versa. Regional uptime is measured and reported independently.

5.3 Coming-soon regions

APAC, SEA, and LATAM regions are not covered by this SLA until they are officially launched. Enterprise customers can request early access with a custom SLA addendum.

6. Exclusions

→Scheduled maintenance during published maintenance windows (48h notice).
→Force majeure events (war, natural disaster, pandemic, government action).
→Third-party service outages outside MnemonX control (e.g. AWS region outage, Twilio SMS outage, Stripe outage).
→The availability, speed, output quality, content rules, pricing or withdrawal of any third-party AI model or its provider (for example OpenAI, Anthropic or Google). We build the interface and the backend that connect you to those models; we do not run them and cannot guarantee them.
→Issues caused by the customer's network, browser, or device.
→Beta features explicitly labelled as experimental.
→Free-plan usage.
→Account or organization suspension due to non-payment on an invoice-billed subscription — not an SLA-qualifying outage.

7. SLA Credits

If monthly uptime falls below the committed level, Enterprise customers are eligible for service credits:

99.0% – 99.4%10% of monthly fee
95.0% – 98.9%25% of monthly fee
< 95.0%50% of monthly fee

Credits must be requested within 30 days of the incident via [email protected]. Credits apply to the affected region only. Maximum credit per month: 50% of monthly fee. Credits are not refunds and have no cash value. For Enterprise accounts, credits are applied to the next invoice rather than refunded.

8. Incident Communication

Status page: Real-time status for both US and EU regions at mnemonx.ai/status. Subscribe to email/SMS alerts for region-specific incident updates.

P1 incident response (Enterprise): Initial response within 2 hours, status updates every 30 minutes, root cause analysis within 5 business days.

About the AI models

When you chat, or ask for text, a picture, a video, a voice or code, your request is answered by an artificial-intelligence model made and run by an independent third-party provider (for example OpenAI, Anthropic, Google or another provider listed in our Privacy Policy). MnemonX builds the interface, your memory wallet, the credits and the backend that connect you to those models. We do not own, train or operate those models and we do not claim them as our own. We are an independent customer of those providers: we pay them to process your requests, and the credits and plan fees you pay MnemonX cover that cost together with the interface, memory and backend we provide. We are not affiliated with or endorsed by the companies that make the models. Product and company names belong to their owners. Which model answers can change over time, a provider can change or withdraw a model or have an outage, and AI output can be wrong, so check it before you rely on it. Read the details in our Terms.

9. Contact

Support: [email protected]

Enterprise SLA / custom addendum: [email protected]

Security incidents: [email protected]

MnemonX · mnemonx.ai

Service Level Agreement