Silogy
SiLogy Technologies, Inc. · Coding
Silogy builds Viv, an on-premise AI agent that debugs failing chip verification tests by reading code, logs, and waveforms, then writes a bug report with the likely root cause.
About Silogy
What Is Silogy
Silogy is the company behind Viv, an AI agent built for digital design verification. The team went through Y Combinator's W24 batch and introduced Viv on Hacker News in March 2024. You hand it a failing test case, along with the code, the log directory, and any waveform dumps, and it works out why the test failed.
The product targets the slowest part of the hardware cycle: the stretch between a red regression and a human figuring out what broke. Viv reads the same artifacts an engineer would, cites the lines of code and the signals it used, and turns that into a written bug report. Silogy says teams cut debug time by upwards of 90%, which is a vendor claim rather than an independent measurement.
Nothing runs in Silogy's cloud. Viv installs inside your own infrastructure, so design data stays on your machines. The only outbound traffic goes to the LLM provider you configure.
Getting Started
- Install Viv on a machine inside your network with the single Node.js package, then confirm it runs from your shell.
- Point Viv at the version of the code used by the failing test, either as directories or as Synopsys VCS style filelists.
- Collect the logs, artifacts, and waveform files for that one failing test into a single directory.
- Add your LLM API key, then submit the test run and read the Markdown bug report it returns.
Product Information
A quick look at Silogy's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Verification engineers
- RTL and design engineers
- Small chip teams without a dedicated debug specialist
Tasks
- Triaging overnight regression failures
- Root-causing one stubborn test
- Writing up a bug for handoff
- Checking a fix
Scenarios
- Shops that can't send IP offsite
- CI/CD pipelines
- Ramping up on an unfamiliar block
Key features
Autonomous Root Cause Analysis
Viv takes a failing test as its starting point and works through the code, logs, specs, and waveforms on its own. It doesn't just flag an error line. It builds a case for where the bug likely lives and cites the artifacts behind that call. Silogy positions the output as a suggestion an engineer reviews, not a verdict.
Waveform Debugging at Multi-Gigabyte Scale
Viv reads FST, VCD, and FSDB dumps and queries signal behavior directly instead of loading everything into a viewer. That style of waveform analysis matters on a long regression with deep traces, where the difference is between a question you can answer and one you skip. You can also ask it about a specific window around the failure.
On-Premise by Design
Everything runs inside your infrastructure: the agent, the analysis, and the artifacts it reads. The only network call out is to the LLM API you set up. For teams under NDA with a foundry or an IP vendor, that constraint usually decides whether a tool is allowed at all.
Bring Your Own LLM
You supply the key for OpenAI, Claude, Llama, or any vLLM-compatible endpoint, and Viv routes its reasoning through that model. That keeps model choice and spend under your control, and it means a provider change doesn't force a tool migration. Teams already running Claude Code can drive Viv through that CLI instead.
Command Line and CI/CD Automation
Viv runs from a small CLI: install the package, then submit a test run with your code and logs attached. When stdin isn't a terminal, it goes headless and prints a Markdown report to stdout. There's an optional VS Code extension if you'd rather read results in an editor.
A Knowledge Base That Grows With Your Project
Viv keeps a knowledge base that updates as it works through your runs. Over time it carries context about your design, your testbench conventions, and past failures, so later debug sessions start with more background than the first one did.
Pros and cons
Pros
- Setup is short: one Node.js package and a submit command, no cluster to stand up first.
- Design data stays on your machines, which clears the usual security review for chip work.
- You pick the LLM and pay that provider directly, so model cost stays visible and reversible.
- Headless mode turns a failing nightly test into a Markdown artifact your CI system can store.
- Reports point at specific code and signals, so a reviewer can check the reasoning instead of trusting it.
Cons
- There's no public price list. You have to book a demo and talk to sales before you know what it costs.
- Viv needs a failing test plus readable code, logs, and waveforms. It won't invent a testbench or fix a setup that never ran.
- Answer quality tracks the model you plug in, and long waveform sessions can run up a real token bill.
- Silogy's own docs say Viv can be wrong, so every report still needs an engineer to confirm it.
Frequently asked questions
Viv is Silogy's AI agent for digital design verification. It takes a failing test case and works out the likely root cause from the code, log files, waveform dumps, and specs, then writes up what it found.
Related content
Explore related tools, skills, and articles for Silogy.
Silogy Alternatives
Base Code
Base44 (Wix) · Coding · ProductivityBase Code is a shared cloud development environment from Base44. Teammates describe changes to an existing GitHub codebase in chat, preview them live and send them back as pull requests.
Google EmbeddingGemma 2
Google DeepMind · Writing · Image · CodingGoogle EmbeddingGemma 2 is a 740M-parameter open model that turns text, code, images, audio, and video frames into vectors in one 768-dimension space, running on phones and laptops.
GitGlow
GitGlow · Coding · ProductivityGitGlow is a local-first desktop Git client that lets independent developers review their own code and their AI agents' changes in one place before shipping.