Hey HN! I'm Daniel, cofounder of GrowthX and Ben's colleague (who posted it). We have about 20 engineers building AI agents and workflows for companies like Lovable, Webflow, Airbyte. Output is the framework we extracted from that work. It runs our AI infrastructure and we open-sourced it.
We kept hitting the same problems: writing and iterating on prompts at scale, orchestrating API calls that fail unpredictably, tracking costs, testing non-deterministic code, building datasets from production data, organizing repos so coding agents perform well. And every piece of tooling was a different SaaS product that didn't talk to the others.
We built Output around three ideas:
1. Make it easy for devs and coding agents to create and modify workflows in one or a few shots.
Filesystem first. Everything your agent needs lives in self-contained folders, full context visible without hunting. TypeScript and Zod provide the first validation layer for whether your workflow is correct.
2. One framework, minimal tooling sprawl.
We got tired of scattering data across SaaS products that don't talk to each other. Prompt files, evals, tracing, cost tracking, credentials all live in one place.
Your data stays on your infrastructure. Under the hood, we built on Temporal for orchestration. It's a hard problem and we weren't going to reinvent the wheel they've perfected. Open source and self-hostable, or Temporal Cloud. We wrapped it so you don't need to learn Temporal upfront, but the full power is there underneath.
3. A flat learning curve.
Our team is web engineers at different levels. We didn't want anyone to learn Python, five different tools, or the nuances of workflow idempotency before they could ship. We baked in conventions: same folder structure, file names, patterns across every workflow. Advanced features like Temporal primitives, evals, LLM-as-a-judge stay out of the way until you reach for them.
We've been building production workflows this way for over a year.
We extracted it, cleaned it up, and wanted to put it in front of people who'd push on it.
Docs and a video building a HN AI digest newsletter from scratch: https://output.ai
I often see people asking how to come up with product ideas, and I think this video is a solid answer (for B2B):
## Here’s a TL;DR of video:
The Two Paths to Starting a Business
First route is identifying and solving an urgent, burning problem that you have in your existing business/job. The alternative path is to align your business model with a significant macro trend.
## AI in SaaS is a macro trend happening now
Past technological shifts can provide ideas how these trends happen and the strategies that can used to navigate it.
## Three Principles
1. Identify the Macro Trend: Recognizing AI as a transformative force in SaaS is the first step.
2. Learn from History: Looking back at how previous platform shifts have played out. Transition from desktop to cloud-based, from cloud to mobile accessibility.
3. Innovate Strategically: Instead of attempting to overhaul existing platforms entirely, focus on innovating specific components or aspects to make them significantly better.
## Five core areas:
1. CRM
2. ERP
3. CMS
4. HCM (Human Capital Management - HR)
5. BI
These are the backbone of the SaaS industry. The introduction of AI is expected to redefine these platforms further.
## How to do it:
Focus on improving functionalities within existing core players in these 5 areas, instead of trying replace. Focus on Land grab and expansion. Aggregate data from existing players and focus on forecasts, automation and replacement. By improving certain aspects, you can position yourself as essential complements to existing systems, gradually expanding their scope. OR, the rapid changing landscape presents opportunities not just for growth and expansion but also for strategic exits through acquisitions by the existing players.
We have noticed that many of our customers are promoting programmers and designers to leadership positions faster than before due to financial constraints. However, many (if not most) companies, especially early-stage ones, lack proper guidance and frameworks to support these newly promoted leaders.
To help those in this situation, we are hosting a free 60-minute webinar on March 22 at 10:00 AM PT. During the session, we will introduce a framework designed to create a leadership operating model specifically for internally promoted managers. We have developed this framework/methodology over the past 8 years while working with over 30,000 people.
I hope this webinar will be helpful to the HN community.
PS: The recordings of our previous webinars covering other leadership related topics is also accessible for free here https://kyt.tips/webinars-hn
I'm not sure about denormalize full-time jobs, but normalizing WFH is a good thing. It does require different practices though.
Our company has a ton of training material on managing remote teams, and due to high demand because of the virus, we decided to make our eBook on "Managing Remote Teams" free. Maybe we can help ease the transition for teams moving to WFH.
The 60+ pages book is the result of months of research and interviews with successful remote companies (it’s usually a part of our paid product). We collected tips on pretty much everything, from onboarding to communication best-practices, to tools you should consider.
We've received a ton of questions on how to manage teams remotely this week as well. We decided to make our eBook on "Managing Remote Teams" free, to help ease the transition for teams moving to WFH.
This 60+ pages book is the result of months of research and interviews with successful remote companies (it’s usually a part of our paid product). We collected tips on pretty much everything, from onboarding to communication best-practices, to tools you should consider.
Many managers don't actually know how to handle WFH well. To help with the transition we decided to give away our eBook on the subject (which is usually paid). You can download the ePub/PDF here: https://knowyourteam.com/m/managing_remote_teams
Maybe share it internally and see if they change the policy. Getting a new job now might be tricky.
We kept hitting the same problems: writing and iterating on prompts at scale, orchestrating API calls that fail unpredictably, tracking costs, testing non-deterministic code, building datasets from production data, organizing repos so coding agents perform well. And every piece of tooling was a different SaaS product that didn't talk to the others.
We built Output around three ideas:
1. Make it easy for devs and coding agents to create and modify workflows in one or a few shots.
Filesystem first. Everything your agent needs lives in self-contained folders, full context visible without hunting. TypeScript and Zod provide the first validation layer for whether your workflow is correct.
2. One framework, minimal tooling sprawl.
We got tired of scattering data across SaaS products that don't talk to each other. Prompt files, evals, tracing, cost tracking, credentials all live in one place.
Your data stays on your infrastructure. Under the hood, we built on Temporal for orchestration. It's a hard problem and we weren't going to reinvent the wheel they've perfected. Open source and self-hostable, or Temporal Cloud. We wrapped it so you don't need to learn Temporal upfront, but the full power is there underneath.
3. A flat learning curve.
Our team is web engineers at different levels. We didn't want anyone to learn Python, five different tools, or the nuances of workflow idempotency before they could ship. We baked in conventions: same folder structure, file names, patterns across every workflow. Advanced features like Temporal primitives, evals, LLM-as-a-judge stay out of the way until you reach for them.
We've been building production workflows this way for over a year.
We extracted it, cleaned it up, and wanted to put it in front of people who'd push on it.
Docs and a video building a HN AI digest newsletter from scratch: https://output.ai
Happy to answer questions.