Data retention policy
Sim implements tiered data retention based on subscription level and data category to balance operational needs with storage efficiency. For free-tier users, workflow execution logs and associated files are retained for 7 days (configurable via environment settings), after which they are automatically processed for archival and deletion. Paid subscription tiers (Pro, Team, Enterprise) benefit from extended retention periods. Account data is maintained throughout the active account lifecycle plus 30 days following a deletion request to allow for account recovery. Third-party API data (such as Google API integrations) is retained during active feature use plus 7 days after access revocation or account deletion. Usage analytics logs are preserved for 90 days for service improvement purposes, with security-related logs retained up to 1 year to support incident investigation. Transaction records are maintained for up to 7 years to satisfy legal and tax compliance requirements.
Data archiving and removal policy
Prior to permanent deletion, execution logs are archived to cloud storage in a structured format (logs/archived/{date}/{id}.json) containing comprehensive metadata including workflow identifiers, execution IDs, timestamps, and archival dates. This archival process ensures data can be retrieved for compliance or audit purposes before final removal. Once archived, the original database records and associated execution files are permanently deleted in batched operations to minimize system impact. Soft deletion is employed for knowledge bases and documents, setting a deletedAt timestamp rather than immediate hard deletion, allowing for recovery within a grace period. Orphaned workflow state snapshots are cleaned up one day after the log retention period expires. Users may request data deletion through in-app settings or by contacting support, and Sim Studio will comply with such requests within the published retention timeframes.
Data storage policy
Sim employs PostgreSQL with the pgvector extension as its primary database (hosted on PlanetScale), enabling both relational data storage and AI-powered vector embeddings for knowledge bases and semantic search capabilities. File storage is handled through a flexible provider system supporting AWS S3, Azure Blob Storage, or local filesystem deployment, with context-specific bucket configurations for knowledge bases, chat attachments, workflow execution files, and user profile assets. All sensitive credentials, including API keys and OAuth tokens, are encrypted using AES-256-GCM encryption with authenticated encryption tags before storage. Data is encrypted in transit using TLS/HTTPS and at rest through provider-native encryption mechanisms. Access is governed by role-based permissions with comprehensive audit logging, and the system implements cascading deletion rules to maintain referential integrity—workspace deletion automatically removes associated workflows, execution logs, schedules, webhooks, and permissions while optionally preserving published templates as orphaned records for continued access.
Data center location(s)
United States
Data hosting details
Our data is hosted on AWS ECS, and we use PlanetScale Postgres as our database provider.
App/service has sub-processors
no
App/service uses large language models (LLM)
yes
LLM model(s) used
GPT-4o/4.1/5/5.1/5.2, O1/O3/O4-mini, Claude Sonnet/Opus/Haiku 4.x/3.x, Gemini 2.x/3.x, Grok 3/4, Deepseek Chat/V3/R1, Mistral/Magistral/Ministral, Codestral, Devstral, Cerebras Llama, Groq models, Azure/Vertex variants, Ollama, vLLM, OpenRouter
LLM retention settings
Execution logs retained 30 days by default, deletable anytime. Data is not used to train models. Self-hosted allows custom retention.
LLM data tenancy policy
Prompts and responses are isolated per workspace with row-level security. No cross-tenant data access is permitted.
LLM data residency policy
Cloud deployments use US-based infrastructure. Self-hosted users choose their region. Third-party LLM providers process data per their own policies.