Skip to main content

Stacks / Solo developer shipping a SaaS

Solo developer shipping a SaaS

An agent-assisted editor or CLI agent, a Postgres backend, auth, billing, hosting, errors and product analytics — only tools with VibeToolRank test scores.

Who it is for

Indie hackers and solo engineers who write code and want a free-friendly stack from idea to first paying user.

Only tools with status tested and a published score appear below. Compare editions on each tool page — some scores use free/BYOK models, not the vendor's top paid model.

1. Write and change code

Goal: Day-to-day coding with an AI editor or terminal agent.

Editors: Cursor and Kiro. Terminal agents: Claude Code (Anthropic Sonnet 5.5 API edition; OpenRouter free edition also on that page), Aider, OpenCode and Cline. Prefer the edition label on each tool page.

Cursor 8.4/10

Cursor 3.25.6 for Linux, Free Hobby plan, Agent with default model Cursor Grok 4.6 Medium

Kiro 9.7/10

kiro-cli 2.37.1, Kiro Free (AWS Builder ID), model auto, agent v3 Default/vibe (not Spec)

Claude Code 9.3/10

Claude Code CLI 2.1.314 with Anthropic Claude Sonnet 5.5 (claude-sonnet-5-5) via API key; OpenRouter free edition also t

Aider 9.3/10

Aider 0.86.2 CLI with OpenRouter free model cohere/north-mini-code:free

OpenCode 9.2/10

OpenCode CLI 1.19.2, free Zen model opencode/big-pickle, no account

Cline 9.2/10

Cline CLI 3.0.72 with OpenRouter free model cohere/north-mini-code:free

OpenAI Codex 7.4/10

Codex CLI 0.163.0 with OpenRouter free model cohere/north-mini-code:free (no ChatGPT login; not OpenAI's models)

2. Catch bad PRs

Goal: Review AI-written diffs before merge.

Scored on our seeded R1 review PR. Sourcery led; Semgrep is stronger as SAST than as a general review bot.

Sourcery 8.8/10

Sourcery GitHub app, Team trial then Open Source plan, benchmark R1

CodeRabbit 8.1/10

CodeRabbit GitHub app, 14-day Advanced trial (no card), benchmark R1

Greptile 7.8/10

Greptile GitHub app, 14-day Pro trial, benchmark R1

Semgrep 8.1/10

Semgrep Community Edition (open-source CLI, registry rules), not Semgrep AppSec Platform

3. Database

Goal: Managed Postgres or reactive backend on a free tier.

Same tested set as the non-coder path.

Neon 9.5/10

Neon Free plan, API v2 + serverless HTTP driver, Postgres 18.6

Supabase 9.2/10

Supabase Cloud, Free plan (ap-southeast-2); self-hosted Docker edition also tested: 8.6

Convex 8.7/10

Open-source local backend (anonymous local deployment via Convex CLI), not Convex Cloud

4. Auth + payments

Goal: Accounts and checkout.

Clerk or Better Auth for identity; Stripe or Polar for billing.

Clerk 8.1/10

Clerk Cloud, Hobby free plan (development instance)

Better Auth 8.2/10

Open-source library (self-hosted), not Better Auth Infrastructure

Stripe 9.2/10

Stripe test mode (free, no card)

Polar 8.8/10

Polar sandbox, REST API v1, organization access token

5. Host + observe

Goal: Ship, watch errors, learn from usage.

Sentry (Developer) and PostHog free tiers cleared our checks for errors and product analytics.

Vercel 9.4/10

Vercel Hobby free plan (no card), CLI 64.3.1

Railway 9.0/10

Railway trial (no card), CLI 5.69.0

Netlify 8.7/10

Netlify Free plan (CLI 27.12.1)

Sentry 9.3/10

Sentry Developer (free) plan, US region, @sentry/node 11.8.0

PostHog 9.0/10

PostHog Cloud US free tier, capture API + HogQL query API

Gaps (not recommended yet)

No tested email/transactional provider yet (Resend, Loops, etc.).
No tested full CRM or support desk.
Claude Code Anthropic Sonnet 5.5 API edition is live (9.3); OpenRouter free edition also 9.3 on the same page. Opus not run.

All stacks · Markdown · Methodology