Show HN: Zenflow – orchestrate coding agents without "you're right" loops (zencoder.ai)
While building our IDE extensions and cloud agents, we ran into the same issue many of you likely face when using coding agents in complex repos: agents getting stuck in loops, apologizing, and wasting time.
We tried to manage this with scripts, but juggling terminal windows and copy-paste prompting was painful. So we built Zenflow, a free desktop tool to orchestrate AI coding workflows.
It handles the things we were missing in standard chat interfaces:
Cross-Model Verification: You can have Codex review Claude’s code, or run them in parallel to see which model handles the specific context better.
Parallel Execution: Run five different approaches on a backlog item simultaneously—mix "Human-in-the-Loop" for hard problems with "YOLO" runs for simple tasks.
Dynamic Workflows: Configured via simple .md files. Agents can actually "rewire" the next steps of the workflow dynamically based on the problem at hand.
Project list/kanban views across all workload
What we learned building this
To tune Zenflow, we ran 100+ experiments across public benchmarks (SWE-Bench-*, T-Bench) and private datasets. Two major takeaways that might interest this community:
Benchmark Saturation: Models are becoming progressively overtrained on all versions of SWE-Bench (even Pro). We found public results are diverging significantly from performance on private datasets. If you are building workflows, you can't rely on public benches.
The "Goldilocks" Workflow: In autonomous mode, heavy multi-step processes often multiply errors rather than fix them. Massive, complex prompt templates look good on paper but fail in practice. The most reliable setups landed in a narrow “Goldilocks” zone of just enough structure without over-orchestration.
The app is free to use and supports Claude Code, Codex, Gemini, and Zencoder.
We’ve been dogfooding this heavily, but I'd love to hear your thoughts on the default workflows and if they fit your mental model for agentic coding.
Download: https://zencoder.ai/zenflow YT flyby: https://www.youtube.com/watch?v=67Ai-klT-B8
24 comments
[ 1.3 ms ] story [ 48.1 ms ] threadThen there's a blurb about the CEO who claims "AI doesn't need better prompts. It needs orchestration." which is something I have always felt to be true, especially after living through highly engineered prompts becoming suddenly useless when conditions change because of how brittle they are.
I might even give this a shot and I usually eschew AI plugins because of how cloud connected they are.
I am a nobody, but I think these people are making a bunch of right moves in this AI space.
nice
First of all kudos for the nice UI. I like when apps looks well. Onboarding process was smooth. I paired it with Zencoder's agent (as mentioned I use their VSCode plugin and already had a sub).
I used it to implement a small refactoring for my side project. What I like compared to plugins, I did not have to switch between agents or explicitly ask to write a plan/spec. It's I guess one of the core ideas behind the app and feels really AI-ish because it's not code editor (similar to claude code). The only thing I missed in the process is rendered markdown for previews. But I did not used the app for long, maybe there is an option to render markdown.
Overall great experience so far. Gonna explore it more. Wanna try it with Gemini and Claude Code. Again kudos it's not locked to use only Zencoder's agents.
Also, I found unexpected use case for it. Even when I need to only change couple lines of code, I just run quick fix workflow for it, because Zenflow automatically creates worktree, branch, commit etc. And PR is created with few clicks. It'll seems like a minor thing, but it irritates me a lot to do all this stuff myself for small changes. One thing I miss here is automatic PR name and description creation according to templates my company uses.
One question: I see this supports custom workflows, which I love and want to try out. Could this support a "Ralph Wiggum"-style [0][1] continuous loop workflow? This is a pattern I've been playing around with, and if I could implement it here with all the other features of this product, that would be pretty awesome.
[0] https://paddo.dev/blog/ralph-wiggum-autonomous-loops/ [1] https://github.com/onorbumbum/ralphio
Create a new task with your prompt, and hit "Create" (instead of "Create and Run"). The interface will show a little hint "Edit steps in plan.md", with 'plan.md' being clickable. Click on it and edit it, experimenting with some ideas. {Bonus tip: toggle "Auto-start steps", to keep it Ralph-y)
I just winged the workflowsbelow, and it worked for the prompt I threw at it. If you like it, you can save it as your custom workflow and use it in the future. If you don't like it - change to your preference.
Now, I prefer a slightly different flow: Implement > Review > [Fix] (and typically limit the loop to 3 times to avoid "divergence"). We'll ship some pre-built templates for that soon. Our researchers are currently working on various variations on our private datasets.
--- # Quick change
## Configuration - *Artifacts Path*: {@artifacts_path} → `.zenflow/tasks/{task_id}`
---
## Agent Instructions
This is a quick change workflow for small or straightforward tasks where all requirements are clear from the task description.
### Your Approach
1. Proceed directly with implementation 2. Make reasonable assumptions when details are unclear 3. Do not ask clarifying questions unless absolutely blocked 4. Focus on getting the task done efficiently
This workflow also works for experiments when the feature is bigger but you don't care about implementation details.
If blocked or uncertain on a critical decision, ask the user for direction.
---
## Workflow Steps
### [ ] Step: Implementation
Implement the task directly based on the task description.
1. Make reasonable assumptions for any unclear details 2. Implement the required changes in the codebase 3. Add and run relevant tests and linters if applicable 4. Perform basic manual verification if applicable
Save a brief summary of what was done to `{@artifacts_path}/report.md` if significant changes were made.
After you are done with the step add another one to `{@artifacts_path}/plan.md` that will describe the next improvement opportunity.
So you got me with the hook, and you bullet three features, but where’s the resolution of the hook issue? You left me with the hook?? What am I missing?
I am just a hobbyist but was curious how you’re thinking through the pricing plans.
One thought - vendors like cursor.ai have the benefit of highly tuned prompts, presumably by programming language, as the result of their user bases. How is it possible to compete with this?
On another note, I have played around with v0 etc, but AFAIK there is no really good UX/UI AI tool that can effectively replace a designer in the way that coding tools are replacing engineers (to a certain extent).