How I built a 2-agent code review loop for local Ollama workflows

Haichi Ops — Mon, 03 Aug 2026 18:54:15 +0000

When running code reviews with local LLMs, a single model can either hallucinate non-existent bugs or generate generic advice you end up ignoring.

To make local AI code review more useful, I built a closed Reviewer vs. Verifier loop for local Ollama workflows.

The Architecture: Two Local Agents, One Loop

Instead of trusting one model's output, the workflow splits the job into two roles:

Agent 1 (Reviewer): Reads the git diff or file changes. It searches specifically for logical flaws, security vulnerabilities, edge cases, or missing unit tests.
Agent 2 (Verifier): Takes the Reviewer's list of findings and actively challenges them. If a finding is weak or unsupported, the Verifier pushes it out of the action list. If it holds up, the next step stays visible.

The goal is not to make the model "always right". The goal is to make weak claims easier to catch before you act on them.

Why Local-First?

Many agent workflows eventually ask you to move private workspace context into somebody else's control plane.

I packaged this workflow into HAICHI, a desktop workspace for Windows and Linux that connects to local Ollama models and keeps the workflow state inspectable.

Key features:

Local-first workflow: Run Reviewer and Verifier style loops around local models.
Visible evidence trail: Keep task, review, challenge, and result in one workspace instead of scattered chat tabs.
Scoped execution: Keep actions bounded to the workflow you explicitly run.
Practical limits: Control how much concurrent agent work runs on your machine.

Try it on your own code

HAICHI Personal is free to try.

Website: https://haichi.app
Supported OS: Windows 10/11, Linux (Ubuntu/Debian/Arch)

If you're already using Ollama for real development work, test the Reviewer vs. Verifier loop on one change and let me know where it helps, where it is too noisy, and what your local setup looks like.

I got tired of juggling Ollama tabs, terminals, and agent chats

Haichi Ops — Mon, 13 Jul 2026 18:55:41 +0000

Running models locally is the easy part. Keeping multiple agents, prompts, plans, and project context organized is where my setup kept falling apart.

I would have Ollama running in one terminal, separate chats in different windows, notes in another file, and no clear answer to a basic question: which agent is doing what, with which model, inside which workspace?

So I built HAICHI, a self-hosted control room for local AI agent workflows on Windows and Linux.

The workflow I wanted

My minimum useful loop was:

Detect the Ollama models already installed on the machine.
Give two agents different roles and model assignments.
Run a plan, discussion, or sequential pipeline without rebuilding the context by hand.
Keep plans, messages, and results in a local SQLite database.
Allow file access only for explicitly selected models and workspace folders.

HAICHI v1.1 now does that from one local dashboard.

What the free build actually includes

The Personal build is free and does not require activation. It includes:

Windows x64 and Linux x64 packages
local Ollama model discovery
up to two local agents
one active workflow at a time
plans, discussions, and pipelines
persistent local state
optional, workspace-scoped file permissions

The limits are deliberate. I want the first run to answer one question: is coordinating two local agents in one workspace more useful than juggling separate chats?

Developer Pro is a separate licensed build for cloud providers, unlimited agents and workflows, and private Tailscale access. You do not need Pro to test the local workflow.

Honest limitations

This is an early desktop product, not a hosted team SaaS.

You need Ollama installed for local models.
Tool reliability depends on the selected model following structured tool instructions.
Smaller models can be useful for chat and planning but may ignore file-tool formats.
File and command permissions are off by default because local automation should not receive broad access silently.

I am looking for three real first workflows

I do not need more generic launch impressions. I need three people who already use Ollama to try one real task and tell me exactly where they stop.

Try HAICHI Personal free: haichi.app

If download, installation, model discovery, or the first workflow blocks you, use the Tell us what blocked you link on the download section. Short and critical feedback is more useful than a polite review.

What is the first two-agent local workflow you would try: parallel research, code review, planning, or something else?

DEV Community: Haichi Ops