Module 10 of 10 · CLI Comparison Track
100%
AI CLI Tools Compared  ·  M10  ·  Final Module

Choosing Your Stack

After 10 modules, the answer is not "pick one." Most professional developers run 2–3 AI CLI tools in complementary roles. This module gives you the decision framework, six developer profiles with concrete recommendations, a proven hybrid workflow, and where each tool is heading in 2026–2027.

Claude Code
Gemini CLI
GitHub Copilot CLI
Section 1

The Key Insight: Not Either/Or

The workshop analogy

BEFORE: A carpenter in 1900 had one multi-purpose tool — a pocketknife. It cut rope, whittled wood, spread butter, and occasionally served as a screwdriver. It did everything; it excelled at nothing.

PAIN: When power tools arrived, the question "should I get a hammer or a saw?" was nonsensical. Professional carpenters owned both — they had different jobs. The hammer was not a substitute for the saw; they were complementary, each doing what the other couldn't.

MAPPING: AI CLI tools are the same. GitHub Copilot is the always-on inline assistant in your IDE — it autocompletes as you type. Gemini CLI is the free-tier workhorse for lightweight automation — PR summaries, issue triage, doc updates. Claude Code is the autonomous agent for hard work — complex refactors, architecture reviews, multi-file changes. Using all three is not excessive; it is professional tooling discipline.

Complementary use cases — where each tool dominates and where they overlap

The typical hybrid stack costs ~$20–35/month per developer and covers every AI workflow: inline completions in the IDE (Copilot), fast free automation in CI (Gemini CLI), and high-quality autonomous coding and architecture review (Claude Code). The tools are not competing for your loyalty — they are competing for different jobs.

Section 2

Interactive Decision Tree

Click any question to expand its answers and follow the path to your recommended stack. Multiple paths are valid — your context determines the right starting point.

AI need $0 / month — no budget at allStart: Gemini CLI (free tier)
BUnder $15/month is fineStart: Claude Code (API, Haiku model)
C$10–20/month per user (team subscription)Start: Copilot Individual/Business
DBudget is not the constraint — quality isHybrid: Claude Code + Gemini CLI + Copilot
AYes — it is the primary use caseGemini CLI (free tier covers most teams) or Claude Code (best quality)
BNice to have, not criticalAny tool works — choose based on other criteria
CI assumed Copilot could do thisCopilot CLI cannot do autonomous CI automation — use Claude Code or Gemini CLI
AYes — Windows PowerShell only, no WSLGemini CLI has the best native Windows support; Claude Code is also good
BWindows + WSL or cross-platformAll three tools work well — choose on other criteria
CmacOS or Linux primarilyAll three tools work equally well on Unix
AYes — quality over everythingClaude Code with claude-sonnet-4-6
BGood enough quality at lowest costGemini CLI with Pro model for complex tasks
CI mainly want inline IDE completionsGitHub Copilot in VS Code or JetBrains
AHeavily — most infrastructure is GoogleGemini CLI primary + Claude Code for code review
BPartially — some Google servicesGemini CLI for Google integrations; Claude Code for complex work
CNo — AWS/Azure/GitHub hostedClaude Code or Copilot depending on IDE priority
Section 3

Six Developer Profiles

Real recommendations for the most common developer archetypes. Find your profile and use it as a starting point — then adjust based on your specific constraints.

Profile 01
Indie Hacker / Solo Dev on a Budget
Gemini CLI (free) Copilot Individual ($10/mo)
Gemini CLI free tier covers all automation and code review needs. Copilot Individual provides IDE completions — the most productive surface for solo builders. Total cost: $10/mo. Upgrade to Claude Code when a project becomes complex enough to need deep reasoning (e.g., a large refactor or security audit).
Profile 02
Startup Team of 5
Claude Code (primary) Gemini CLI (CI/free tasks)
Claude Code for all architecture, PR review, and complex work (~$30–35/mo total API). Gemini CLI free tier for CI automation (PR summaries, issue labeling). No Copilot subscription needed — Claude Code covers the code review use case and Gemini covers completions-in-CI. Total: ~$30–40/mo for the whole team.
Profile 03
Enterprise Team (Compliance-first)
Copilot Business (compliance) Claude Code (architecture review)
Copilot Business provides SOC 2, audit logs, IP indemnification, and GitHub Enterprise integration — features enterprise procurement requires. Add Claude Code for high-value architecture and security reviews where quality matters most. Gemini CLI optional for Google Cloud shops. Total: ~$27/user/mo (Copilot Business + Claude API est.).
Profile 04
Windows-First Dev Shop
Gemini CLI (best Windows UX) Claude Code (secondary)
Gemini CLI was built with Windows PowerShell as a first-class citizen — the Plan Mode and GEMINI.md experience translate cleanly to Windows workflows. Claude Code also supports PowerShell well. Skip Copilot CLI if you need autonomous work; use Copilot in VS Code for completions. Primary total: $0–10/mo.
Profile 05
AI-Heavy Automation / DevOps
Claude Code (autonomous work) Gemini Plan Mode
For teams whose primary AI use case is autonomous infrastructure changes, deployment scripts, and code generation pipelines: Claude Code is the strongest autonomous agent. Gemini CLI Plan Mode is useful for reviewing complex multi-step workflows before execution. Copilot CLI's interactive-only model makes it unsuitable. Estimated: $50–150/mo for a 5-person ops team.
Profile 06
Google Ecosystem Shop
Gemini CLI (primary) Workspace integration Claude Code (review)
Teams running on GCP, BigQuery, and Google Workspace get native integrations through Gemini CLI's @google/mcp-workspace extension — Docs, Sheets, Gmail access out of the box. Use Claude Code for code review where reasoning depth matters. The combination costs $0–8/mo for most Google-ecosystem teams using free tier for Gemini.
Section 4

The Hybrid Workflow

Here is the real-world hybrid workflow that a professional developer team uses, with each tool doing what it does best. Total monthly cost for this full stack: ~$27–35/developer.

IDE Completions
Copilot in VS Code
Complex Refactor
Claude Code
PR Review
Claude Code (CI)
Issue Triage
Gemini CLI (free)
Changelog Draft
Gemini Flash (free)
Architecture ADR
Claude Code (Sonnet)
Monthly cost breakdown for the hybrid stack — per developer
The $35/mo full stack is cheaper than one junior developer's hour

At a fully-loaded developer rate of $100/hr, the full hybrid stack (~$35/mo) costs less than 21 minutes of developer time per month. Saving even 15 minutes per day (conservatively) generates $500/mo in value — a 14x return. The question is not whether to invest in AI tooling; it is which combination maximizes your team's output for their specific workflow.

Section 5

Where Each Tool Is Going: 2026–2027

The AI CLI landscape is moving fast. Here is where credible public roadmaps and announced features point for each tool:

Claude Code Roadmap
  • Computer use: control browser + desktop autonomously
  • Multi-agent orchestration (subagents) for parallel tasks
  • Deeper GitHub integration: branch → PR → merge cycle
  • Claude Opus 5 / next-gen models with longer context
  • Built-in memory across sessions (persistent CLAUDE.md)
  • Enterprise policy engine for large org deployments
Gemini CLI Roadmap
  • Project Antigravity migration (full IDE replacement preview)
  • Gemini 2.5 Ultra in CLI with 2M token context
  • Native Android / mobile development integration
  • Deeper Workspace: real-time Docs/Sheets co-editing
  • Firebase integration for full-stack app generation
  • Video / multimodal input for UI bug reports
Copilot Roadmap
  • Copilot Workspace: autonomous PR creation (web preview)
  • GitHub Models: any model in Copilot context
  • Multi-file agents in VS Code (Copilot Edits)
  • Copilot CLI expansion: beyond shell commands to code review
  • GitHub Actions native Copilot (not just copilot-action)
  • Integration with GitHub Issues and Projects
The convergence is real — but not yet

All three tools are converging on similar capabilities: autonomous multi-file editing, CI integration, and large context awareness. By late 2027, the feature differentiation may narrow significantly. In the near term (2026), the gap remains: Claude Code is the strongest autonomous agent, Gemini CLI is the most cost-efficient, and Copilot is the IDE completion leader. Build your stack on current reality and plan to re-evaluate in 12 months.

Section 6

Course Complete

You've Completed AI CLI Tools Compared

Across 10 modules, you've mastered the three major AI CLI tools — their architecture, configuration, integration patterns, cost models, and when to use each. You can now make informed, defensible tooling decisions for any team size and context.

Claude Code configuration Gemini CLI Plan Mode Copilot CLI limits MCP integrations CI/CD automation TCO analysis Hybrid stack design Prompt engineering Security patterns Agentic workflows
Start the Capstone Project

The capstone asks you to build the same URL Shortener microservice three times — once with Claude Code, once with Gemini CLI, once with GitHub Copilot CLI — documenting the experience at each step. This hands-on comparison will cement the abstract knowledge from the course with real experience.

What you can do now

You can evaluate AI CLI tools systematically, configure MCP servers for all three tools, build automated PR review workflows in GitHub Actions, calculate TCO for your team, design a hybrid stack optimized for your specific constraints, and explain the trade-offs to technical leadership and procurement. These are skills that few developers in the industry have today — and that distinction is commercially valuable.

Final Knowledge Check

Five questions drawing on the full course. These test integrated understanding, not just individual module facts.

Question 1 of 5
A developer says "I'll just use whichever AI tool my teammates use — it's all the same." Based on what you've learned across this course, what is the best response?
Correct. Tool consistency within a team has real value — shared CLAUDE.md files, shared GEMINI.md configs, shared GitHub Actions workflows. But "they're all the same" ignores: (1) Copilot CLI cannot do autonomous CI automation; (2) Claude Code charges per token while Copilot charges per seat; (3) Gemini CLI has a free tier that saves money at small scale; (4) MCP configuration is different for each tool; (5) Plan Mode is unique to Gemini CLI. Using the same tool as teammates for shared config is reasonable; believing the tools are interchangeable leads to wrong expectations.
Question 2 of 5
A startup wants to add automated PR review to their GitHub Actions pipeline for free, configure MCP servers for database access, AND get IDE code completion. Which combination achieves all three?
Correct. This is the hybrid stack at work. Gemini CLI's free tier covers CI automation (PR review, issue triage) and supports MCP server configuration via GEMINI.md — zero cost. GitHub Copilot Individual ($10/mo) provides best-in-class IDE code completion in VS Code. Claude Code could also handle CI and MCP, but it costs money (API tokens). Copilot CLI cannot do autonomous CI automation or MCP. The Gemini + Copilot combination covers all three requirements at minimum cost ($10/mo).
Question 3 of 5
What is the single most important factor that makes Claude Code better suited for complex autonomous work (multi-file refactors, security audits) than Gemini CLI or Copilot CLI?
Correct. The key differentiator is the autonomous agent loop combined with the model's reasoning depth. Claude Code doesn't just generate code — it reads your codebase (via file tools), writes changes, runs tests (via bash tool), reads the test output, and fixes failures — iterating until the task succeeds or it determines the approach is wrong. This multi-step agentic loop, combined with claude-sonnet-4-6's ability to reason about complex architectural trade-offs, is what makes it superior for complex tasks. Gemini CLI has a similar tool-use loop (Plan Mode), but the model's reasoning on complex architectural problems currently lags Sonnet. Copilot CLI doesn't have a tool-use loop at all.
Question 4 of 5
An enterprise team of 50 developers wants to add Claude Code to their stack. Their security team asks: "who sees our code when Claude runs?" What is the correct answer for Claude Code with an Anthropic Enterprise contract?
Correct. Anthropic's API, at the Business and Enterprise tiers, offers a zero-data-retention option: prompts (including your code) are not stored after the response is generated and are not used for model training. The practical answer for a security team is: with a signed Anthropic Enterprise agreement and the zero-retention option enabled, your code is processed in transit but not persisted. This is the same model that large banks and healthcare organizations use for LLM API access. The key contract terms to verify: data processing addendum (DPA), zero-retention clause, and the data residency region. All three tools have similar options at enterprise tiers.
Question 5 of 5
After completing this course, a developer wants to recommend a tool to their team lead. Which framing makes the strongest case for adopting AI CLI tooling?
Correct. The strongest business case combines: (1) concrete cost numbers (the $10–35/mo range you can now cite precisely), (2) specific ROI calculation tied to developer rates that managers understand, and (3) named, specific automation tasks with estimated time savings. Vague appeals to productivity feelings or peer pressure don't survive procurement review. But "we'll eliminate 1hr/week of PR review wait time per developer at 10 developers = 10hr/week = $1,000/week saved, at a cost of $300/month" — that calculation closes budget decisions. The course has given you the exact numbers and examples needed to make this argument precisely. Use them.