The life agentic

Which tool should I use?

If GitHub student benefits work for you, start with GitHub Copilot CLI.

If you want a paid general-purpose coding agent, start with Claude Code. Then move to OpenCode when you want more provider choice and control, and use llm when you want scripting or Python integration.

If you are unsure how you actually get access to a model, read Model access first.

If you specifically want a local or offline route, read Local models and Ollama next.

Intended learning outcomes covered on this page

After working through this page, students should be better able to:

• distinguish between an LLM tool, an underlying model, and a model-access route
• choose an appropriate access route based on availability, cost, and desired workflow
• choose an appropriate tool for a task by distinguishing between coding-agent workflows and scripting or pipeline workflows
• explain why repository-grounded terminal agents support different kinds of work than ordinary browser chat
• explain when local models via Ollama are a good fit and when hosted tools are the better choice

Short answer

• use GitHub Copilot CLI (copilot) if you want the simplest student setup
• use Claude Code if you want one strong paid coding tool
• use opencode if you want a more capable open source coding agent
• use llm if you want repeatable prompts, shell pipelines, or Python scripts

Comparison

Tool	Best when	Main strength	Main limitation
GitHub Copilot CLI (copilot)	you want to get started quickly with GitHub student benefits	simplest onboarding and strong GitHub integration	less flexible than the more configurable tools
Claude Code (claude)	you want one strong paid coding tool for repo work	polished workflow, strong repository help, skills and CLAUDE.md	paid subscription and less open or configurable than OpenCode
OpenCode (opencode)	you want longer, more agentic coding sessions in a repository	stronger project workflow, more providers, more control	more setup and more power means more care is needed
Python package llm	you want scripts, structured output, or Python integration	best fit for automation and reproducible workflows	not primarily a full coding agent

Suggested path for this course

1. If GitHub student benefits are available, start with GitHub Copilot CLI.
2. If you want a paid fallback or a single paid coding tool, start with Claude Code.
3. Move to OpenCode when you want a more configurable open source repo workflow.
4. Learn llm when you want reproducible command-line or Python workflows.

Common situations

• “I just got GitHub student benefits and want something that works quickly.” Use GitHub Copilot CLI.
• “I want one paid tool and do not want to depend on GitHub approval first.” Use Claude Code.
• “I want an agent to inspect a repository, plan a change, and edit several files.” Use OpenCode.
• “I want to put LLM use inside a shell pipeline or a Python script.” Use llm.
• “I want structured JSON output or a script I can rerun later.” Use llm.
• “I want an open source terminal agent that can use GitHub Copilot or other providers.” Use OpenCode.
• “I want local or offline workflows with open-weight models.” Read Local models and Ollama, then use OpenCode or llm.

The important difference

The biggest difference is not just model quality. It is workflow.

Browser chat often depends on copy-paste and your own description of the project. Terminal agents can inspect files and use tools directly, which is why they are usually much more effective for repository work.

• GitHub Copilot CLI is the easiest entry point.
• Claude Code is the strongest polished paid general coding tool.
• OpenCode is the strongest choice for interactive agentic coding in a project.
• llm is the best choice when you want repeatable prompts and programming access.

If you want to understand why the coding-agent tools feel similar in some ways and different in others, read Agentic concepts. That page reconnects planning, context windows, subagents, and memory across the main coding-agent tools, then uses llm as a contrast case.

Self-check

Try to answer these before looking at the suggested answers:

1. You want the quickest path from GitHub student benefits to a working terminal agent. Which tool is the best first choice?
2. You want one paid general-purpose coding tool for repository work and do not want to wait on GitHub approval. Which tool fits best?
3. You want to inspect a repository, plan a multi-file change, and edit files interactively. Which tool fits best?
4. You want a repeatable shell pipeline or Python script rather than an interactive coding agent. Which tool fits best?

Suggested answers:

1. GitHub Copilot CLI.
2. Claude Code.
3. OpenCode.
4. llm.

Three important things that are not tools

• Model access explains how you actually reach a model.
• Which model should I use? explains which model fits which kind of task once you have a tool and a way to reach models.
• Skills let you reuse playbooks or procedures with an agent.
• AGENTS.md gives agents project instructions in a standard place.

Pages in this section

• Register with GitHub as a student
• Understand model access: GitHub, Claude, APIs, Zen, and Ollama
• Learn when local models and Ollama make sense
• Get started with Claude Code
• Get started with GitHub Copilot CLI
• Get started with OpenCode
• Understand the shared agentic concepts
• Learn what skills are
• Learn what AGENTS.md is for
• Learn a reusable verification workflow
• Get started with the Python package llm
• Which model should I use?
• Problematic cases of using AI

This site is open source. Improve this page.