Tadashi Shigeoka

OpenAI Codex has become an essential part of our development workflow. With our AGENTS.md providing project context, Codex understands our monorepo architecture and coding conventions, autonomously handling implementations that would take hours manually. Code reviews are faster and more thorough as Codex catches edge cases and potential issues we might miss. The ability to run tasks in parallel and work autonomously in the background has dramatically increased our development velocity. Thank you for building such a powerful coding agent!

Claude Code has transformed how we build Giselle. It understands our monorepo architecture, coding conventions, and domain concepts — making implementation feel like pair programming with a senior engineer who knows the entire codebase. Code reviews that used to take hours now take minutes as Claude spots edge cases and suggests improvements. Our development velocity has never been higher. Thank you for building the future of software development!

Sentry keeps Giselle reliable by catching every error across our entire stack — server, edge, and client. When something goes wrong in a complex AI workflow, we know about it immediately with full context: user info, team data, and detailed traces. The Next.js integration is seamless, and the ability to attach rich context to errors makes debugging production issues so much faster. Thank you for helping us ship with confidence!

Google's Gemini models bring unmatched context, multimodal, and image generation capabilities to Giselle. With a 1M token context window, our workflows can process entire codebases, long documents, and even video content. From Gemini 2.5 Flash Lite for fast free-tier tasks, to Gemini 2.5 Flash and Pro for balanced and advanced reasoning, Gemini 3 Pro and Flash for the most demanding agentic workflows, and Gemini 2.5 Flash Image for native image generation — we leverage the full Gemini family. Google Search Grounding keeps responses current, while native multimodal support handles text, images, audio, and video seamlessly. Thank you for pushing the boundaries of AI capability!

OpenAI's GPT models power a wide range of Giselle's AI workflows. From GPT-5 nano for quick, cost-effective tasks on our free tier, to GPT-5 mini for balanced everyday use, GPT-5 and GPT-5.1 Codex for complex reasoning and coding agents, and GPT-5.2 for the most demanding challenges — we leverage the full GPT family. The Reasoning Effort parameter lets users fine-tune the balance between speed and depth, while Web Search keeps workflows grounded in real-time information. Thank you for building such a versatile AI platform!

Claude powers the intelligence behind Giselle's AI workflows. We leverage the full Claude family — Haiku for fast, cost-effective tasks on our free tier, Sonnet for balanced everyday workflows, and Opus for the most demanding reasoning challenges. Extended Thinking lets our workflows reason through multi-step problems, while multimodal capabilities enable understanding PDFs and images. The API's reliability and Claude's unique ability to follow complex instructions make it perfect for orchestrating sophisticated AI workflows. Thank you for building AI that truly thinks!

GitHub is at the heart of everything we do — not just as our development platform, but deeply integrated into Giselle itself. Our AI workflows can be triggered by GitHub events (issues, PRs, discussions), use 30+ GitHub tools for AI agents to manage repositories, and even build semantic search indexes over entire codebases. From development to production, GitHub enables us to create truly automated software engineering workflows. Thank you for building the platform that makes developer automation possible!

PostHog gives us deep insights into how users interact with our AI workflow editor. Understanding user journeys through complex visual workflows — where they click, what features they use, where they get stuck — helps us continuously improve the experience. The self-hostable, open-source nature aligns perfectly with our values. Thank you for building such a powerful and privacy-conscious analytics platform!

Stripe powers our entire billing infrastructure. The new Billing for LLM tokens feature (Private Preview) made it incredibly easy to implement usage-based pricing for AI API consumption — what would have taken weeks to build ourselves was ready in days. Combined with Pricing Plan Subscriptions for our team plans, Stripe handles all the complexity of hybrid billing so we can focus on building the best AI workflow experience. Thank you for staying ahead of the AI era!

Langfuse gives us complete visibility into our AI workflows. From tracking token usage and costs across multiple LLM providers to tracing every generation and embedding operation, it's become essential for understanding how our users interact with AI. The tagging and session tracking features make debugging and optimization so much easier. Thank you for building such a powerful observability platform for LLMs!

Trigger.dev powers our AI workflow execution engine. Running complex multi-step AI workflows reliably in the background — with support for long-running tasks and automatic retries — would have been incredibly complex to build ourselves. Thank you for making background job processing so developer-friendly!

Supabase powers our entire backend infrastructure — from seamless user authentication to reliable file storage and secure secrets management with Vault. Supabase Storage was a game-changer for our AI workflow engine's reliability. Thank you for building such a developer-friendly platform!

Giselle is built with the help of many coding agents. Reviewing, refining, and extending what those agents generate—and collaborating with them to shape something that didn’t exist before—requires an editor that is truly fast. Without Zed’s speed and responsiveness, Giselle’s release would have taken much longer. We’re very grateful for it.

As a designer on the team, Cursor was surprisingly one of the most impactful tools while building Giselle. It lowered the friction between “I have a design intention” and “this is how it’s actually implemented.” Being able to read, tweak, and question real code in-context helped me participate earlier and more concretely in technical decisions — not just at the handoff stage. It didn’t turn me into an engineer, but it made collaboration much tighter.

We couldn't have built Giselle without Vercel. Next.js is our rock-solid foundation. The Vercel AI SDK brilliantly powers our multi-model orchestration across Claude, GPT, and Gemini. AI Gateway has been a game-changer for stable, reliable model access. Vercel's world-class ecosystem made our open-source AI workflow platform a reality. Thank you!