Foreman Subprocessors
Foreman is an AI orchestration platform that serves as an AI Chief of Staff, offering studio, CRM, campaigns, prospecting, and publishing features.
- Subprocessors
- 27
- Locations
- 6
- Categories
- 13
- Last scanned
| Name | Category | Purpose | Location | Data processed |
|---|---|---|---|---|
| Descope | Authentication | Authentication, magic link, and social login |
 United States
|
Identity attributes, login metadata, authentication events |
| GitHub | Developer Tools | Sign-in via GitHub OAuth when a user chooses that login method; tracking product bug reports and feature requests submitted through in-app feedback |
United States
|
Identity attributes (GitHub username, profile, and email) shared during OAuth exchange; text of in-app feedback reports, page URLs, and automated triage summaries |
| Neon | Database | PostgreSQL database hosting |
United States
|
Workspace configurations, user accounts, agent memory, operational logs |
| Stripe | Payment Processing | Payment processing |
United States
|
Billing details, subscription state, invoice history |
| Vercel | Cloud Infrastructure | AI Gateway — routing and transport for Foreman-managed LLM inference, and a transit conduit for customer-key (BYOK) inference | United States (provider-managed processing region not guaranteed) | Prompts, conversation context, and model responses in transit; customer provider API keys passed per request (never stored by Foreman at Vercel) when BYOK traffic routes through the gateway |
| OpenAI | AI | LLM inference and orchestration; safety/abuse screening (content moderation); failover speech transcription and speech synthesis | United States (provider-managed processing region not guaranteed) | Prompts, business context, tool execution results (transiently); dictation audio and read-aloud message text when a speech failover rung engages (transiently) |
| DeepInfra | AI | LLM inference (open weights) and managed text-to-speech synthesis (read-aloud); from November 25, 2026, text embeddings and reranking for semantic search and duplicate-record detection | United States (provider-managed processing region not guaranteed) | Prompts, conversation excerpts (transiently); assistant message text sent for read-aloud speech synthesis (transiently); CRM record, message and document text sent for embedding or reranking (transiently; never for PHI-Local workspaces) |
| OpenRouter | AI | LLM model routing and fallback | United States (provider-managed processing region not guaranteed) | Prompts, payloads requested during capability escalations (transiently) |
| Kilo | AI | LLM inference gateway — the primary managed serving lane for models not available on the Vercel AI Gateway | United States (provider-managed processing region not guaranteed) | Prompts, conversation context, and model responses (transiently) |
| Groq | AI | Speech-to-text transcription of user dictation audio (managed speech recognition) | United States (provider-managed processing region not guaranteed) | Voice audio recordings from dictation, sent per request for transcription (transiently) |
| AssemblyAI | AI | Speech-to-text transcription with speaker labels for conversations recorded in the Foreman desktop app; failover transcription of voice dictation | United States (provider-managed) | Audio of conversations chosen to be recorded (microphone, and for calls the other participants' audio), fetched by AssemblyAI from the workspace's own host for transcription; the resulting transcript. Voice dictation audio, sent per request (transiently), only when earlier speech providers have failed |
| TypeSafe AI | AI | Decision-model inference (Jev) — classifying requests, inbound conversations and planned tasks for routing and triage. Off until a workspace owner or admin turns it on through an explicit consent prompt. | United States (provider-managed processing region not guaranteed) | The message or conversation text being classified, with recent conversation context, or a planned task's description, its objective and the titles of the tasks planned with it, sent directly to TypeSafe AI (not retained: zero data retention) |
| Braintrust | AI | LLM observability (run telemetry on the hosted edition) |
United States
|
Per-run operational metadata only: workspace identifier, agent id, model/provider, token counts, cost, and latency. Prompt and output content is sent only if the operator explicitly enables payload logging; the local/desktop edition sends nothing. |
| Nylas | Managed email and calendar connectivity for the Hub (message sync, sending, and calendar access), and the Meeting Notetaker that records online meetings a user sends it to |
United States
|
Email messages and bodies, calendar events, contact and participant addresses, and OAuth grants for connected mailboxes and calendars; for the Meeting Notetaker, the audio of the Zoom, Google Meet or Microsoft Teams meeting, participants' meeting display names, and the speaker-labelled transcript | |
| Composio | Integration | Third-party integrations, tool execution, and managed web search for agent research |
United States
|
OAuth tokens, execution payloads between agents and external platforms, and agent search queries (transiently) |
| Pipedream | Automation | Third-party integration connection and execution |
United States
|
OAuth metadata, execution payloads, and integration logs where enabled |
| Fly.io | Cloud Infrastructure | Application hosting and runtime infrastructure, and — when you choose Foreman cloud hosting — the machine and data volume your workspace's execution host runs on |
United States
|
User requests, application responses, and transient runtime data; for a cloud-hosted workspace, its execution database and runtime data (messages, CRM records, work products) on a volume provisioned for that workspace |
| Hetzner | Cloud Infrastructure | Virtual-machine infrastructure for optional Foreman-managed hosted workspaces (your workspace execution host, when you choose Foreman cloud hosting) | United States regions (provider is Germany-based Foreman provisions U.S. regions) | The hosted workspace's execution database and runtime data (messages, CRM records, work products) on VMs provisioned for your workspace |
| Contabo | Cloud Infrastructure | Virtual-machine infrastructure for optional Foreman-managed hosted workspaces | United States regions (provider is Germany-based Foreman provisions U.S. regions) | The hosted workspace's execution database and runtime data on VMs provisioned for your workspace |
| DigitalOcean | Cloud Infrastructure | Virtual-machine infrastructure for optional Foreman-managed hosted workspaces; DNS management for hosted domains |
United States
|
The hosted workspace's execution database and runtime data on VMs provisioned for your workspace; hostnames and DNS records |
| GoDaddy | Hosting | DNS management for hosted domains |
United States
|
Hostnames and DNS records for hosted workspaces; no customer content |
| Cloudflare | Cloud Infrastructure | R2 private object storage; secure tunnel transit for remote access to your own device |
United States
|
Uploaded documents, attachments, and export bundles (R2); encrypted remote-device traffic in transit (Tunnel) |
| Firecrawl | AI | Managed web-retrieval backend (Foreman-keyed) for agent research, page scraping, and structured extraction |
United States
|
Agent search queries, specified URLs, and retrieved page content (transiently) |
| Resend | Transactional email delivery (sign-in, beta invitations, account and system notifications) |
United States
|
Recipient email addresses and the contents of transactional messages | |
| ipapi.co | Data Enrichment | Fallback IP geolocation for sanctions/jurisdiction screening at sign-up and billing, used only when the primary on-device GeoIP lookup cannot resolve an address |
United States
|
Client IP address (used transiently to resolve approximate country/region; only on a local-lookup miss) |
| U.S. Department of Commerce — Consolidated Screening List API (trade.gov) | Compliance | Sanctions and denied-party screening required for legal compliance |
United States
|
Names submitted for screening against U.S. government sanctions lists |
| Plaid | Fintech | Bank-account connection and financial-data aggregation for the Finance features (importing the accounts, balances, and transactions you direct us to sync) |
United States
|
Bank-account connection tokens and the account, balance, and transaction data you direct Foreman to import; the bank you link is a Customer-Directed Service |
No subprocessors match your filters.
No changes recorded yet. Changes will appear here after the next scan detects updates.
Summary
27
Total Subprocessors
6
Locations
13
Categories
63%
in United States
Subprocessors by Country
Subprocessors by Category
Biggest overlap with
Other companies using the most of the same subprocessors
Common in your industry
Top subprocessors used by other AI / SaaS companies