crewai-tokportal

CrewAI integration for TokPortal.

TokPortal is the managed social infrastructure API: real TikTok, Instagram and YouTube accounts created, warmed and operated by human account managers in 16+ countries — exposed as a REST API and an MCP server. No OAuth per account, no 25-posts/day cap, no app review.

Two ways to give your crew TokPortal super-powers:

Installation

pip install crewai-tokportal
# for the MCP variants:
pip install "crewai-tokportal[mcp]"     # == crewai-tools[mcp]

Get an API key at https://app.tokportal.com/developer:

export TOKPORTAL_API_KEY=sk_...

Crew example (native tools)

from crewai import Agent, Crew, Task

from crewai_tokportal import tokportal_tools

operator = Agent(
    role="TokPortal operator",
    goal="Order and schedule managed social accounts for the marketing team",
    backstory=(
        "You operate TokPortal, a managed social infrastructure API. You always check the "
        "credit balance first, create a bundle, configure its videos, then publish it so a "
        "human account manager can start."
    ),
    tools=tokportal_tools(),          # reads TOKPORTAL_API_KEY; or tokportal_tools("sk_...")
    llm="gpt-4.1",
    verbose=True,
)

order = Task(
    description=(
        "Order one new TikTok account in the US with 3 video slots and advanced warming on "
        "'home workout', 'protein snacks', 'gym motivation'. Publish the bundle and report "
        "its id and status."
    ),
    expected_output="The bundle id, its status, and the remaining credit balance.",
    agent=operator,
)

crew = Crew(agents=[operator], tasks=[order])
print(crew.kickoff())

Every tool returns a JSON string: the TokPortal data envelope, or {"error": {"status", "code", "message", "request_id", "retryable"}} so the agent can self-correct.

Full MCP tool set

Recommended: Agent(mcps=[...]) (CrewAI ≥ 1.x)

from crewai import Agent
from crewai_tokportal import tokportal_mcp_server

agent = Agent(
    role="TokPortal operator",
    goal="...",
    backstory="...",
    mcps=[tokportal_mcp_server()],   # MCPServerHTTP, streamable HTTP + Bearer auth
    llm="gpt-4.1",
)

tokportal_mcp_server() accepts MCPServerHTTP kwargs such as tool_filter= or cache_tools_list=True.

Legacy: MCPServerAdapter context manager

from crewai import Agent
from crewai_tools import MCPServerAdapter
from crewai_tokportal import tokportal_mcp_adapter

with MCPServerAdapter(tokportal_mcp_adapter(), connect_timeout=60) as mcp_tools:
    agent = Agent(role="...", goal="...", backstory="...", tools=mcp_tools, llm="gpt-4.1")
    ...

tokportal_mcp_adapter() returns {"url": "https://app.tokportal.com/api/ext/mcp", "transport": "streamable-http", "headers": {"Authorization": "Bearer sk_..."}}.

Tool reference (native)

Individual classes are exported too (TokPortalCreateBundle(api_key=...), …).

Resources

• API docs: https://developers.tokportal.com
• OpenAPI: https://developers.tokportal.com/openapi.json
• MCP server: https://developers.tokportal.com/mcp
• Python SDK: https://github.com/tokportal/tokportal-python
• LangChain package: https://github.com/tokportal/langchain-tokportal
• Support: team@tokportal.com

License

MIT