SessionMart

SessionMart is a metered, pay-as-you-go platform for coding agents. Instead of buying separate subscriptions for Claude Code, Codex, Kimi Code, and other agent CLIs, you install one macOS or Windows client and pay only for the sessions you actually use. It targets developers who want full agent environments (read repos, run commands, edit code) without managing multiple subscriptions, API keys, or endpoints.

Key Features

• One app, multiple agents: Claude Code, Codex, Kimi Code, and Grok Build sessions run from a single client; switch models with a click, no endpoint or key configuration.
• Pay per use, not per subscription: Metered billing with transparent input/output/cache rates shown before you connect, plus a hard budget cap that stops automatically at your limit.
• Sessions, not API keys: You enter the agent UI directly instead of receiving a key to configure yourself, so it works like the official native environment.
• $5 free credit: New accounts get $5 of usage credit to start the first agent session immediately.
• Cross-platform: Native macOS and Windows clients.

Use Cases

Who Should Use This Tool?

• Developers hit by account bans: If your region cannot easily subscribe to Anthropic or OpenAI, SessionMart hosts sessions in supported regions so you can keep using Claude Code and Codex.
• Casual or intermittent users: Pay per session instead of a monthly plan when you only run agents occasionally.
• Multi-agent tinkerers: Try Claude Code, Codex, and Kimi Code side by side without buying each subscription.

Problems It Solves

1. Subscription waste: Fixed monthly plans are expensive when usage is sporadic; metered billing matches spend to actual use.
2. Regional access pain: Agent sessions hosted in supported regions bypass local subscription restrictions.
3. Configuration sprawl: No endpoints, keys, or client parameters to manage; switch models inside the app.

Pricing

Advantages & Unique Selling Points

Compared to Competitors:

1. Metered by use: Unlike Claude Pro or Codex subscriptions, you pay only for sessions you run.
2. Sessions over keys: Other relay services hand you an API key; SessionMart gives you the full agent interface, so it behaves like the official client.
3. Budget hard stop: Rate transparency plus an automatic cap means no surprise overcharges.

What Makes It Stand Out:

• One client for Claude Code, Codex, Kimi Code, and Grok Build.
• Zero-config model switching.
• Daily free credits via Twitter and feedback rewards.

Getting Started

1. Download the macOS or Windows client from sessionmart.ai.
2. Sign up and receive $5 of free credit.
3. Open a session: Pick Claude Code, Codex, or Kimi Code and start working on your repo.
4. Set a budget cap to prevent overruns, then monitor usage in-app.

Frequently Asked Questions

Do I need my own subscription or API key?

No. Sessions are provided by the platform; bring-your-own-subscription is planned but not yet available.

Which agents are supported?

Codex, Kimi Code, and Grok Build today, with Claude Code in progress.

Can I use my own API key?

Not yet. The platform hosts the sessions; "bring and share your own subscription" is under development.

Alternatives

If SessionMart isn't the right fit, consider these alternatives:

• Pi: Coding agent harness with flexible model routing.
• Claude Code: Anthropic's official CLI coding agent.
• Codex: OpenAI's coding agent CLI.
• OpenRouter: Unified LLM API router for key-based access.

Tips & Best Practices

1. Set the budget cap before starting long tasks so spending never exceeds your comfort zone.
2. Watch the rate table before each session; cache read/write rates differ from fresh input.
3. Collect free credit from the Twitter account and in-app feedback rewards to offset small experiments.

Conclusion

SessionMart turns coding agent access into a utility: one client, multiple agents, pay only for what you use. For developers who want Claude Code and Codex without per-product subscriptions or regional hassle, it is a practical metered alternative.