llama-index-tools-instanode

LlamaIndex tool spec for instanode.dev. Lets a LlamaIndex agent provision ephemeral Postgres databases + webhook receivers mid-task. No Docker, no account for the free tier.

pip install llama-index-tools-instanode

Usage

from llama_index.agent.openai import OpenAIAgent
from llama_index_tools_instanode import InstanodeToolSpec

tool_spec = InstanodeToolSpec()
agent = OpenAIAgent.from_tools(tool_spec.to_tool_list(), verbose=True)

agent.chat(
    "Stand up a Postgres DB with pgvector, then store this embedding in it: "
    "<embedding>."
)

The agent gets four tools:

• provision_postgres(name) → postgres:// DSN (pgvector pre-installed).
• provision_webhook(name) → HTTPS receiver URL.
• list_resources() → enumerate resources owned by the current API key.

MongoDB, Redis/cache, NATS queue, and heartbeat-monitor tools are on the roadmap, gated on the matching backend endpoints landing. They live on the feature/full-api branch.

Scoping to specific tools

tool_list = [
    t for t in tool_spec.to_tool_list()
    if t.metadata.name in {"provision_postgres"}
]

Paid-tier credentials

Set INSTANODE_API_KEY in your environment, or pass it explicitly:

tool_spec = InstanodeToolSpec(api_key="sk_...")

Tier model

Related

• Python SDK: https://pypi.org/project/instanode/
• MCP server (for Claude Code/Cursor): https://www.npmjs.com/package/@instanode/mcp
• LangChain variant: https://pypi.org/project/langchain-instanode/
• HTTP API reference: https://instanode.dev/llms.txt

License

MIT.