What Claude Code does
Claude Code is Anthropic's coding agent. It reads a codebase, edits files across a project, runs commands and tests, and works with git to stage changes, write commit messages, create branches and open pull requests. You describe a task in plain language, set the permission level, and review what it proposes.
- Builds features and fixes bugs across multiple files, from a pasted error or a described symptom.
- Explains unfamiliar codebases and runs long refactors and migrations.
- Turns GitHub or GitLab issues into pull requests, and can run code review in GitHub Actions or GitLab CI/CD.
- Connects to outside tools through MCP, and can be customised with a CLAUDE.md instructions file, skills and hooks.
- Runs several agents in parallel and on a schedule, and can be scripted from the command line.
It runs in a terminal (macOS, Linux, Windows), in VS Code and JetBrains, in a desktop app, in the browser at claude.ai/code, on the Claude mobile apps, and from Slack.
Pricing
Claude Code is included in paid Claude plans, or billed per token through a Claude Console (API) account. Monthly prices below are from Anthropic's pricing page, before tax.
| Plan | USD per month | SGD per month | Main limit |
|---|---|---|---|
| Free | 0 | 0 | Limited usage; Claude Code not included |
| Pro | 20 (17 if billed annually) | 25.80 (21.93) | 1x usage every 5 hours and every week |
| Max | 100 | 129.00 | 5x Pro usage |
| Ultra | 200 | 258.00 | 10x Pro usage weekly, no 5 hour limit for verified accounts |
| Team, Standard seat | 25 per seat (20 if annual) | 32.25 (25.80) | More usage than Pro; 2 to 150 seats |
| Team, Premium seat | 125 per seat (100 if annual) | 161.25 (129.00) | 5x a Standard seat |
| Enterprise | 20 per seat, annual, plus usage at API rates | 25.80 plus usage | 20 seat minimum; admins set spend limits |
Console accounts pay standard API token prices instead of a plan fee. Anthropic's cost guide gives an average of about US$13 per developer per active day, or US$150 to US$250 per developer per month, across enterprise deployments (below US$30 per active day for 90% of users). Conversion uses US$1 = S$1.29 (mid-market rate of 1.2896, 19 November 2026, rounded).
Who it suits
- Developers and small engineering teams who want an assistant that works inside their own repository and terminal.
- Agencies and product teams maintaining several codebases that need tests, migrations and pull requests handled faster.
- Technical founders who can read and review code, and want to build and change an application themselves.
Limits to know
- The Free plan does not include Claude Code. Pro, Max, Ultra, Team, Enterprise or a Console account is needed.
- Subscription plans have usage windows that reset every five hours and weekly. When you reach one, you wait for the reset or add paid usage credits.
- On Team and Enterprise plans the seat allowance is shared with Claude chat and Cowork.
- Anthropic's commercial terms say the customer owns outputs and that Anthropic may not train models on customer content from its services. These are the terms for commercial customers.
- The commercial terms bar using the service to build a competing product, train competing AI models or resell it without approval.
- Under the commercial terms, the services and outputs are provided "as is" and "as available" without warranty.
- Agent teams are off by default and use roughly seven times the tokens of a standard session when teammates run in plan mode.
Where a custom build fits
Claude Code is enough when you have developers, or a technical owner, who can review its changes and you want to speed up work on code you own. A business outgrows it when the need is a finished product or an agent that serves customers and staff, with hosting, security and support behind it. For that, see our web application development service.