[!IMPORTANT] CALL-E can place real outbound phone calls. Integrations must plan first and only start a real call when the user clearly intends to place that call. Agent-facing flows should prefer calle call start so execution confirmation data stays inside the CLI.

🚀 Quick install

For most users, the simplest path is to ask your agent to install the portable CALL-E skill.

Copy this stable prompt into a local agent that can run shell commands or install skills:

Install CALL-E for me: https://raw.githubusercontent.com/CALLE-AI/call-e-integrations/main/docs/install/CALL-E-installation-guide.md

The linked guide contains the install steps, so the prompt can stay unchanged when those steps need to evolve.

New users get 20 free calls to get started.

If the prompt flow is not supported in your client, use the manual install guide. It covers Codex, Claude Code, Cursor, OpenClaw, Hermes Agent, CLI-only, and MCP-only setup. The native Claude Code plugin path uses /plugin marketplace add https://github.com/CALLE-AI/call-e-integrations.git#@call-e/claude-plugin@latest.

Install guides: Manual · CLI · Codex · skills.sh · Claude Code · Cursor MCP · Cursor plugin · OpenClaw source

🧯 Troubleshooting

If installation, authentication, or MCP tool verification fails, start with the CALL-E troubleshooting guide.

It covers local agent environment issues such as Cursor sandbox/network restrictions, CONNECT tunnel failed, response 403, calle auth login failures, and verification that plan_call, run_call, and get_call_run are available.

🧠 What CALL-E gives your agent

plan_call → run_call → get_call_run

Capability	Surface
Brokered browser login	calle auth login
Private local token cache	~/.calle-mcp/cli
MCP client configuration	Streamable HTTP URL or calle mcp config
Tool discovery	calle mcp tools
Call planning	calle call plan or MCP plan_call
Real outbound call execution	calle call start, calle call run, or MCP run_call
Status, activity, summary, details, transcript	calle call status or MCP get_call_run
Agent-client UX	$calle in Codex; calle through skills.sh; /calle:calle in Claude Code; calle MCP/skill in Cursor; Phone Call - CALL-E in OpenClaw; ClawHub Prompt flow in Hermes Agent

plan_call creates the call plan and returns run credentials. run_call starts the real outbound call. get_call_run returns progress and final results when available.

⚡ CLI command map

calle auth login
calle auth login --start-only --no-browser-open
calle auth status
calle auth logout

calle mcp config
calle mcp tools
calle mcp call plan_call --args-json '{"to_phones":["+15551234567"],"goal":"Confirm the appointment"}'

calle call plan --to-phone +15551234567 --goal "Confirm the appointment"
calle call start --to-phone +15551234567 --goal "Confirm the appointment"
calle call run --plan-id <plan_id> --confirm-token <confirm_token>
calle call status --run-id <run_id>

Defaults used by the CLI:

Base URL:      https://seleven-mcp-sg.airudder.com
MCP channel:   openagent_oauth
MCP URL:       https://seleven-mcp-sg.airudder.com/mcp/openagent_oauth
Token cache:   ~/.calle-mcp/cli

All command output is JSON except --help. Access tokens are read from the local cache and are never printed.

🛡️ Agent safety contract

• Real calls may contact external people or businesses.
• Always plan first.
• If the user asked to place a call from an agent-facing flow, prefer calle call start so planning and execution confirmation stay inside the CLI.
• If the user only asked to verify setup or draft a plan, do not place the call.
• Do not guess phone numbers, country codes, language, region, execution confirmation data, or run_id.
• Do not print, request, or expose access tokens.

📦 Developer docs

Path	Package	Role
packages/core	@call-e/core	Shared runtime helpers for brokered auth, private token cache files, JSON HTTP, and MCP Streamable HTTP calls.
packages/cli	@call-e/cli	Shared calle command for auth, MCP config, tool listing, and phone-call workflow shortcuts.
packages/codex-plugin	@call-e/codex-plugin	Codex plugin bundle that exposes CALL-E as the $calle skill.
packages/claude-plugin	@call-e/claude-plugin	Claude Code plugin bundle that exposes CALL-E through the shared CLI and /calle:calle.
packages/cursor-plugin	@call-e/cursor-plugin	Cursor plugin bundle that exposes CALL-E through MCP, a calle skill, and a safety rule.
packages/openclaw-cli-skill	@call-e/openclaw-cli-skill	Private validation/source package for the OpenClaw skill published through ClawHub.
packages/skills-sh-skill	@call-e/skills-sh-skill	Private validation package for the portable skills.sh calle skill.

.agents/plugins/marketplace.json              # Codex marketplace entry
.claude-plugin/marketplace.json               # Claude Code marketplace entry
.cursor-plugin/marketplace.json               # Cursor marketplace entry
packages/codex-plugin/plugin/                 # Codex plugin source
packages/claude-plugin/plugin/                # Claude Code plugin source
packages/cursor-plugin/plugin/                # Cursor plugin source
packages/openclaw-cli-skill/skills/            # OpenClaw skill source
skills/calle/                                  # skills.sh compatible skill source
packages/cli/                                  # Shared calle CLI
packages/core/                                 # Shared runtime helpers
examples/                                      # Runnable MCP demos, not an SDK

Stable identifiers

Surface	Identifier	Value
Codex marketplace	name	call-e-codex
Codex marketplace	display name	CALL-E
Codex plugin entry	name	calle
Codex plugin source	path	./packages/codex-plugin/plugin
Claude Code marketplace	name	call-e-claude
Claude Code plugin entry	name	calle
Claude Code plugin source	path	./packages/claude-plugin/plugin
Claude Code CLI attribution	env	claude/claude_code_plugin/<version>
Cursor marketplace	name	call-e-cursor
Cursor plugin entry	name	calle
Cursor plugin display name	displayName	CALL-E
Cursor plugin source	path	./packages/cursor-plugin/plugin
Cursor MCP server	key	calle
Cursor CLI attribution	env	cursor/cursor_plugin/<version>
OpenClaw skill	slug	phone-call-calle
OpenClaw skill	name	Phone Call - CALL-E
skills.sh skill	name	calle
skills.sh skill source	path	skills/calle
skills.sh CLI attribution	env	skills_sh/skills_sh_skill/<version>

🧪 Examples

Runnable MCP client demos live under examples:

• Standard MCP OAuth clients: TypeScript and Python clients for standard MCP OAuth over Streamable HTTP.
• CALL-E broker login MCP clients: TypeScript and Python clients for CALL-E brokered login, local token cache, and MCP HTTP calls.

These examples are runnable demos, not a CALL-E SDK or supported application API.

🧭 Boundaries

• The CLI is not an OAuth server and not an MCP server. It is a local wrapper over the CALL-E broker API and remote MCP HTTP endpoint.
• Codex, Claude Code, OpenClaw, skills.sh, and Hermes Agent integrations intentionally reuse the shared calle CLI for authentication, token caching, JSON output, MCP tool discovery, and call workflow shortcuts.
• The Cursor plugin reuses the existing remote CALL-E MCP server and does not implement a new local MCP server or backend.
• The OpenClaw route in this repository does not register OpenClaw-native tools and does not require a gateway restart from this repository.
• Hermes Agent support uses the ClawHub Prompt flow for the published OpenClaw skill; this repository does not maintain a separate Hermes plugin.
• Future Copilot, VS Code, Gemini, Windsurf, Zed, Cline, Roo, Continue, or other ecosystem integrations should add their own ecosystem-specific entry point instead of sharing the Codex, Claude Code, or Cursor marketplaces.

See docs/agent-integration-layout.md for layout and marketplace naming rules.

🔒 Telemetry / usage data

The calle CLI sends best-effort usage telemetry to CALL-E to diagnose installation, authentication, MCP tool availability, and early usage drop-off before a first plan_call reaches the server.

CLI telemetry includes an anonymous installation ID, CLI version, integration source such as cli/cli/<version>, codex/codex_plugin/<version>, claude/claude_code_plugin/<version>, cursor/cursor_plugin/<version>, openclaw/openclaw_cli_skill/<version>, or skills_sh/skills_sh_skill/<version>, command stage, outcome, error type, and server host/hash. It does not include phone numbers, call goals, OAuth tokens, broker login URLs, full argument JSON, transcripts, or contact data.

Disable CLI telemetry with any of:

DO_NOT_TRACK=1 calle auth status
CALLE_TELEMETRY=0 calle auth status
calle auth status --no-telemetry

Broker and MCP requests still create service-side security, audit, and business operation logs needed to authenticate users and run calls.

👩‍💻 Development

This repository uses Node >=22, pnpm >=10.18.3, Changesets, and GitHub Actions.

pnpm install
pnpm check
pnpm test
pnpm pack:dry-run

Package-specific checks:

pnpm --filter @call-e/core check
pnpm --filter @call-e/core test
pnpm --filter @call-e/cli check
pnpm --filter @call-e/cli test
pnpm --filter @call-e/codex-plugin check
pnpm --filter @call-e/codex-plugin test
pnpm --filter @call-e/claude-plugin check
pnpm --filter @call-e/claude-plugin test
pnpm --filter @call-e/cursor-plugin check
pnpm --filter @call-e/cursor-plugin test
pnpm --filter @call-e/openclaw-cli-skill check
pnpm --filter @call-e/openclaw-cli-skill test
pnpm --filter @call-e/skills-sh-skill check
pnpm --filter @call-e/skills-sh-skill test
pnpm run check:examples

For user-visible package changes, add a changeset. The release workflow publishes changed @call-e/* packages to npm and maintains the @call-e/codex-plugin@latest and @call-e/claude-plugin@latest install aliases.

💬 Community

• Website: heycall-e.com
• Discord: discord.gg/6AbXUzUV8w