Record and Replay on CodeInstances.com

CodeInstances Forks a Run at Any Step, to Run On Live or Try a What-If Answer

Sun, 18 Oct 2026 00:00:00 +0000

CodeInstances v0.6.0 adds fork. Replay runs a program against the whole recording. Fork runs it against the recording up to a step you pick, then lets it carry on from there. You get the past exactly as it happened, and a new future.

Two uses pull this in. The first is picking up where a run went wrong. An agent ran for twenty minutes, called a dozen APIs, and on call nine it got an answer that sent it the wrong way. You don’t want to spend another twenty minutes and another dozen calls getting back to call nine. codeinstances fork -at 9 run.instance answers calls one to eight from the recording in a moment, then sends call nine for real, against today’s servers.

CodeInstances Launches as an Open Source Tool That Keeps Program Runs and Replays Them

Sun, 18 Oct 2026 00:00:00 +0000

CodeInstances v0.1.0 is out. It’s a small command-line tool that keeps a run of any program as a file and replays it later, with the network replaced by what was recorded. It’s open source under the Apache License 2.0 and comes as a single static binary for Linux, macOS and Windows.

The problem it goes after is an old one. Code is kept with great care, in version control, reviewed and tagged. Runs are thrown away. A test fails in CI, someone hits rerun, it passes, and whatever went wrong is gone along with the evidence. An API answers something strange once and never again. A job breaks on Tuesday, and by Wednesday the data it read has moved on.

CodeInstances Now Compares Two Runs and Shows Where the Failing One Went Wrong

Sun, 18 Oct 2026 00:00:00 +0000

CodeInstances v0.2.0 is out, and it adds the third verb. The first release could keep a run and replay it. This one compares two of them: codeinstances diff passing.instance failing.instance.

The usual case goes like this. A job works on Monday and fails on Tuesday, or passes on your laptop and fails in CI. You have both runs as instance files, because you put codeinstances record in front of the command. Now you want to know what was different, and you’d rather not read two logs side by side and guess.

How CodeInstances Was Tested Before Its First Release

Sun, 18 Oct 2026 00:00:00 +0000

A tool that claims to replay runs exactly has to be tested on exactly that. So most of the testing for v0.1.0 does the same thing over and over in different ways: record a program against real servers, shut the servers down, replay it, and check the replay came out the same. Then change something and check the replay notices.

Every number in this post comes from the result files in the code repository, in test/results. They were recorded on October 18, 2026, at commit c2a5975, on a Linux machine with an Intel Xeon processor at 2.10GHz and 2 CPUs, with Go 1.24.7. One script, scripts/record-tests.sh, runs everything and writes the files, so anyone can run it again.

How to Replay the AI Agent Run That Did Something Odd Last Night

Sun, 18 Oct 2026 00:00:00 +0000

An agent ran overnight on a batch of tasks. In the morning one result is wrong. It filed a ticket under the wrong customer, or summarised a document it never fetched, or called a tool twice. You run the agent on the same task to see what happened, and it does the job perfectly.

That’s normal for agents. The model’s answers vary from run to run, so the agent picks different tools in a different order and ends up somewhere else. Running it again doesn’t show you last night’s run. It shows you a new one. With an ordinary program, rerunning is how you debug. With an agent, the only copy of the run that went wrong is whatever you kept at the time.

What a Code Instance Is and Why Program Runs Should Be Kept Like Code

Sun, 18 Oct 2026 00:00:00 +0000

If you’ve written object-oriented code, you already know the idea. A class is a blueprint. Each object you make from it is an instance of that class, with its own state. One Order class, thousands of orders.

Your program works the same way. The code is the blueprint, and every time it runs, that run is an instance of it. Same code, different day, different answers from the servers it talks to, different clock, different random numbers. Each run is its own thing.

Why Replay Matters for AI Agents That Take a Different Path Every Run

Sun, 18 Oct 2026 00:00:00 +0000

Most programs are boring in a good way. Give them the same input and they do the same thing. If one fails, you can usually run it again and watch it fail again.

An AI agent isn’t like that. The model’s answer varies from run to run, so the agent picks different tools, calls them in a different order, reads different pages and ends somewhere else. Ask it the same thing three times and you may get three different runs, and only one of them may go wrong. Run it again to see what happened, and you’re watching a different run.