Data Processing Agreement Subprocessors
This page lists all subprocessors that process personal data on behalf of Abundly AI AB. We maintain this list to provide transparency about our data processing activities and to help our customers comply with their own data protection obligations.
Core Subprocessors
The following subprocessors are always active and are neither LLM- nor capability-specific.
| Subprocessor | Purpose | Location | Data Processed |
|---|---|---|---|
| Vercel | Web application hosting (frontend) | Sweden | Portal application data |
| Google Cloud Platform | Platform hosting (compute, storage, networking) | EU | Application data |
| Google Workspace | Customer support | EU | Customer data shared with us in support cases, e.g. screenshots, conversation logs, files and attachments |
| MongoDB Atlas | Platform database | Sweden | Application data |
| Mailgun (Sinch) | Transactional and system email delivery | EU | Email content and recipient data |
LLM Subprocessors
The following LLM subprocessors are only active when using the corresponding LLM models on our platform.
| Subprocessor | Purpose | Location | Data Processed |
|---|---|---|---|
| Amazon Web Services | Amazon Bedrock-hosted Claude LLM inference (Bedrock Runtime invoke API), speech synthesis (Amazon Polly), speech-to-text (Amazon Transcribe), and live voice (Amazon Nova Sonic) | EU | User queries, prompts, conversation history, and system instructions submitted to AI models via API; text submitted for speech synthesis; audio and video recordings submitted for transcription; voice audio in live voice sessions |
| Anthropic | Claude LLM inference | Global | User queries, prompts, conversation history, and system instructions submitted to AI models via API |
| Google Cloud Platform (Vertex AI) | Gemini LLM inference | EU/Global (selectable) | User queries, prompts, conversation history and system instructions sent to AI models via API |
| OpenAI | LLM inference, embeddings, images, text-to-speech, and voice (Realtime) | Global | User queries, prompts, conversation history, and system instructions submitted to AI models via API |
| Scaleway | Scaleway-hosted LLM inference (Qwen and other model families) | France | User queries, prompts, conversation history, and system instructions submitted to AI models via API |
| xAI | Grok LLM inference | Global | User queries, prompts, conversation history, and system instructions submitted to AI models via API |
Capability-Specific Subprocessors
The following subprocessors are only active when users have enabled specific capabilities on our platform.
| Subprocessor | Purpose | Capability | Location | Data Processed |
|---|---|---|---|---|
| ConvertAPI | Document and PDF conversion, including converting documents for transcription | PDF Generation | Lithuania | Document content and converted files |
| Daytona | Virtual machine sandboxes | Computer Use, Virtual Machine | EU | Code, files, and data uploaded to or generated within agent-provisioned sandboxes |
| ElevenLabs | Speech-to-text (audio transcription) and text-to-speech | Text to Speech | US | Audio and video recordings submitted for transcription; text submitted for speech synthesis |
| Perplexity | Web search and deep research API | Deep Research, Web Search | US | Research queries and results |
| SideGuide Technologies (Firecrawl) | Web page scraping | Web scraping | US | URLs and scraped web content |
| Telnyx | SMS and voice telephony | Make Phone Call, Receive Phone Call, Receive SMS, Send SMS | US | Message content and recipient data |
Last updated: 2026-10-29