Parslee Subprocessors
June 4, 2026
A "subprocessor" is a third party Parslee engages to process customer data on our behalf in order to deliver the Services. This page lists the subprocessors we currently engage, the purpose of each engagement, and the categories of customer data each one receives.
We hold each subprocessor to confidentiality, security, and data-processing obligations consistent with our Privacy Policy and applicable US data-protection law, including the CCPA, and any Data Processing Addendum we have with you.
If you have questions about any item on this list, or how a specific subprocessor handles your data, please contact us at privacy@parslee.ai.
Subprocessors
Microsoft Azure Cosmos DB
- Purpose: Primary application datastore (organizations, accounts, AI Employees, knowledge bases, conversation history)
- Data category: All customer data
- Region: East US 2 (United States)
Microsoft Azure Container Apps
- Purpose: Application hosting for the Parslee web app and backend API
- Data category: Application memory only; customer data passes through during request processing and is not persisted at this layer
- Region: East US 2 (United States)
Microsoft Azure OpenAI Service
- Purpose: Large language model inference for AI Employee responses, agent reasoning, and document generation
- Data category: Prompts and completions (inputs to AI Employees and the responses they generate)
- Region: East US 2 (United States)
Microsoft Azure Application Insights
- Purpose: Application performance monitoring and product telemetry
- Data category: Funnel events and diagnostic metadata, scrubbed at the SDK layer; no message bodies, no attachments
- Region: East US 2 (United States)
Microsoft Azure Key Vault
- Purpose: Secure storage of secrets and cryptographic keys used by the platform
- Data category: Internal credentials only; no customer data
- Region: East US 2 (United States)
Microsoft Azure Communication Services
- Purpose: Voice calling, SMS delivery, and call transcription for voice-enabled AI Employees
- Data category: Call audio, transcripts, phone numbers
- Region: East US 2 (United States)
Microsoft Graph
- Purpose: Microsoft 365 integration (Outlook, Teams, SharePoint, OneDrive) when a customer connects their tenant
- Data category: Customer-authorized M365 data (emails, calendar entries, files, Teams messages); the scope is determined by the permissions the customer grants during connection
- Region: Determined per tenant by the customer's Microsoft 365 data-residency configuration
Sentry (parslee-ai organization)
- Purpose: Engineering error monitoring and crash reporting
- Data category: Stack traces, account identifier, organization identifier, and scrubbed diagnostic metadata. We strip message bodies, request bodies, email addresses, and API keys before transmission
- Region: United States
Stripe
- Purpose: Subscription billing, token-pack purchases, payment processing, and customer-portal management
- Data category: Billing identifiers, payment-method metadata (handled by Stripe directly; Parslee does not see card numbers), email address on the billing contact, organization identifier, and subscription / invoice records
- Region: United States
HubSpot
- Purpose: Customer relationship management for sales-led products (AI Employees, Trainer, ODI, Ink, VoiceRail, ACS, Hydra) and inbound contact records
- Data category: Business contact information (name, email, company, role), conversation records with Parslee sales, and integration metadata when a customer connects their HubSpot account to a Parslee Employee
- Region: United States
ElevenLabs
- Purpose: Text-to-speech voice synthesis, music and sound-effect generation, and speech-to-text transcription for voice-enabled features (Studio, VoiceRail)
- Data category: Text inputs for speech synthesis; audio inputs for transcription; the resulting audio outputs
- Region: United States
Google (Vertex AI / Veo)
- Purpose: Video generation via the Veo model family for Studio
- Data category: Prompts submitted for video generation; the resulting video outputs
- Region: United States
PiAPI
- Purpose: Routing layer for third-party generative-AI models used by Studio (Seedance video generation, Kling lip-sync, image generation)
- Data category: Prompts and reference media submitted for generation; the resulting generated media
- Region: United States
Black Forest Labs (Flux)
- Purpose: Image generation via the Flux model family for Studio
- Data category: Prompts submitted for image generation; the resulting image outputs
- Region: United States
Tavily
- Purpose: Real-time web research used by AI Employees when a research-capable channel is active
- Data category: The search query an AI Employee issues on the user's behalf
- Region: United States
Mixpanel
- Purpose: Server-side product analytics covering feature usage and funnel events
- Data category: Hashed user and organization identifiers, event names, and event properties scrubbed of message bodies and free-text inputs
- Region: United States
DocuSign
- Purpose: Electronic signature collection when an AI Employee initiates a signing workflow on the user's behalf
- Data category: Documents submitted for signature, signer email addresses, and signing-event metadata
- Region: United States
Linear
- Purpose: Issue and project tracking when a customer connects their Linear workspace to a Parslee Employee
- Data category: Customer-authorized Linear data (issues, projects, comments) within the scope the customer grants during connection
- Region: United States
Twilio
- Purpose: SMS delivery and short-code messaging for AI Employees with telephony-enabled channels
- Data category: Phone numbers, message content, and delivery-status metadata
- Region: United States
Vapi
- Purpose: Voice-agent configuration and call-routing layer used by VoiceRail
- Data category: Call-session metadata, configuration parameters, and call audio routed through VoiceRail
- Region: United States
Zoom (Meeting SDK)
- Purpose: Meeting join, audio capture, and participation for voice-enabled AI Employees that attend Zoom meetings
- Data category: Meeting audio, participant metadata, and meeting-context metadata for the meetings a customer authorizes
- Region: United States
Microsoft Bot Framework
- Purpose: Channel connectivity for AI Employees deployed into Microsoft Teams and other Bot-Framework-supported surfaces
- Data category: Message content, channel identifiers, and conversation-routing metadata for the channels the customer authorizes
- Region: Determined per channel by Microsoft's regional routing (commonly United States or European traffic-manager endpoints for global delivery)
Microsoft Power BI
- Purpose: Analytics and reporting integrations when a customer connects their Power BI workspace to a Parslee Employee
- Data category: Customer-authorized Power BI data (datasets, reports, dashboards) within the scope the customer grants during connection
- Region: Determined per tenant by the customer's Microsoft 365 data-residency configuration
Apple iCloud (CalDAV)
- Purpose: Calendar synchronization when a customer connects an iCloud calendar to a Parslee Employee
- Data category: Customer-authorized calendar entries within the scope the customer grants during connection
- Region: United States
Infrastructure providers (not subprocessors)
The following providers support our infrastructure but do not process customer data. We list them here for transparency.
- GitHub — source control and continuous integration. Stores Parslee's application source code; does not process customer data.
Updates to this list
We will update this page when we add, remove, or materially change a subprocessor engagement. Each update will be reflected in the "last updated" date at the top of this page. Customers entitled to advance notice of new subprocessors under their Data Processing Addendum will receive that notice through the contact on file before the change takes effect.
Questions or contractual records
Email privacy@parslee.ai. On request and where contractually permitted, we will share copies of subprocessor security attestations (SOC 2 Type II, ISO 27001, executed DPAs) that our subprocessors make available to us.