MCP tools catalog
For setup, see the MCP setup guide. The server exposes Streamable HTTP at https://api.10ex.ai/mcp and stdio for local Claude Desktop and Cursor.
Tools are grouped by workflow stage so an AI agent can do an end-to-end motion without falling off a cliff.
A0. Onboarding and account
| Tool | Description |
|---|---|
signup | Create a new account and company; returns bootstrap JWT and first API key |
verify_email | Verify a freshly minted account |
login | Email and password to JWT |
whoami | Current user, company, role, plan |
get_company / update_company | Company profile |
list_team_members / invite_teammate | Workspace membership |
list_api_keys / mint_api_key / revoke_api_key | API key self-service |
get_billing_status | Plan, usage caps, remaining credits (read-only) |
A. Setup and enablers
| Tool | Description |
|---|---|
list_connectors | Status for Gmail, Outlook, Webflow, Sanity, Google/LinkedIn/Meta Ads, HubSpot, Salesforce |
start_connector_oauth | Returns OAuth URL for a provider |
get_connector_status | Poll an in-flight OAuth grant |
list_brand_knowledge / upsert_brand_knowledge | Brand context CRUD |
ingest_url_into_knowledge | Scrape a URL into brand knowledge |
B. Audience and CRM
| Tool | Description |
|---|---|
create_lead, bulk_import_leads, search_leads, enrich_lead | Leads CRUD |
create_segment, add_leads_to_segment, list_segments | Segments |
create_company, search_companies | Companies |
create_form, list_forms | Capture forms |
C. Agent Store discovery
| Tool | Description |
|---|---|
list_agents_in_store | Catalog of hireable agents (verbatim slugs from the seed) |
get_agent_details | Connectors, knowledge, sub-agents, I/O schema for one slug |
list_active_agents | What the workspace is currently running |
D. Email Sequence Launch (flagship)
| Tool | Description |
|---|---|
preview_email_sequence | Generate a sequence without sending, for human approval |
approve_email_sequence | Release a previewed sequence to send |
launch_email_sequence | One-shot generate and send (when approval_required: false) |
get_email_sequence_status | Per-recipient send/open/reply state |
send_single_outreach | Single-email cold outreach via email-outreach-sender |
E. Other channel launches
| Tool | Description |
|---|---|
generate_blog | Brief to outline to draft to hero image, via blog-generation |
publish_blog | Push to Webflow or Sanity |
generate_ad_copy | LinkedIn, Meta, Google variants |
launch_ad_campaign | Wraps google-ads-manager, linkedin-ads-manager, meta-ads-manager |
generate_webinar | Outline, script, slides, avatar, stitched video |
run_company_research | Wraps company-research |
run_market_research | Wraps market-research-crew |
run_growth_audit | Wraps growth-audit |
F. Workflows and orchestration
| Tool | Description |
|---|---|
list_workflow_templates | Template Gallery |
create_workflow_from_template | Instantiate a template |
trigger_workflow | Manual run |
G. Docs and introspection
| Tool | Description |
|---|---|
search_docs | Qdrant-backed semantic lookup over this docs corpus |
get_setup_checklist | For a goal, return the ordered prerequisite tools to call |
How to chain tool calls without a hardcoded checklist
Every “launch” tool returns a Pydantic model with the resulting entity ID and a next_actions array, so an agent can chain calls without a setup checklist on the client side. If you want one anyway, get_setup_checklist returns an ordered prerequisite list for a stated goal.
What scopes and permissions look like
v1 keys carry scope
['*']. Per-tool scopes are tech debt #20 in10ex_fastapi/TECH_DEBTS.md.
Last updated on