plmbr/notebook-intelligence — A JupyterLab extension supporting Claude Code, Copilot, Ollama, and OpenAI-compatible LLMs, with MCP…

Description and use

**Brief Summary (2‑3 sentences)**
plmbr/notebook‑intelligence is a JupyterLab extension that brings together Claude Code, Copilot, Ollama and any OpenAI‑compatible LLMs into a unified orchestration layer, offering MCP‑based skills, plugins, and notebook agents. It lets you turn ad‑hoc prompts and tool calls into repeatable, memory‑backed agent workflows that can be coordinated across multiple LLMs. With 317 stars, recent commits, and a Python‑centric codebase, it is positioned as a production‑ready OSS candidate for AI‑enhanced notebook environments.

**Value**

• **From isolated prompts to reusable agents** – The extension abstracts LLM calls, tool usage, and memory handling into “agents” that can be versioned, shared, and invoked programmatically, eliminating the need to rewrite prompt logic for each notebook.
• **Multi‑agent orchestration** – By supporting several LLM back‑ends and an MCP (Multi‑Component‑Pipeline) model, teams can route tasks to the most suitable model (e.g., Claude for reasoning, Copilot for code generation) and chain their outputs automatically.
• **Standardised tool‑use pipelines** – Built‑in plugins let notebooks call external services (APIs, databases, CLI tools) in a consistent way, making data‑driven research reproducible and auditable.

**Practical Adoption Path**

1. **Install & enable** – Add the extension via pip install plmbr-notebook-intelligence and enable it in JupyterLab (jupyter labextension install).
2. **Configure LLM back‑ends** – Supply API keys or local endpoints (Ollama, Claude, OpenAI) in the provided config.yaml.
3. **Define agents** – Use the UI or a small Python DSL to declare agents, their skills, and memory scopes (e.g., per‑project or per‑notebook).
4. **Integrate into notebooks** – Replace raw openai.ChatCompletion calls with agent.run(prompt); existing notebooks continue to work while gaining orchestration benefits.
5. **Scale & version** – Export agent definitions as YAML/JSON, store them in a repo, and reuse across teams or CI pipelines.

**Production Readiness**

• **Activity & community** – 317 stars, 59 forks, recent commits (as of 2026‑06‑25), and active issue discussions indicate a healthy maintainer base.
• **Stability** – The core API (API/SDK/CLI) is stable, with clear versioning and extensive type hints in Python, reducing integration risk.
• **Ecosystem fit** – Works natively with JupyterLab, the de‑facto environment for data science, and can be called from CI/CD pipelines via its CLI.
• **Risks to address** – Final due‑diligence on the license (MIT‑style but verify), security posture of bundled plugins, and confirmation of long‑term maintainers is still required.

Overall, plmbr/notebook‑intelligence offers a mature, low‑friction way to embed LLM‑driven agents into notebook workflows and is ready for pilot deployments in production‑grade data‑science environments.

Problem

plmbr/notebook-intelligence helps turn isolated prompts and tools into repeatable agent workflows.

Outcome

High for an OSS candidate: recent activity, adoption, and ecosystem signals are strong enough for a serious pilot.

Workflow

• coordinate multi-agent workflows
• add tool-use pipelines
• standardize agent memory

Evidence

A JupyterLab extension supporting Claude Code, Copilot, Ollama, and OpenAI-compatible LLMs, with MCP, skills, plugins, and notebook agents.

Source