Discover AI Applications

Browse applications built on modern technologies. Explore PoC and MVP applications created by our community and discover innovative use cases for modern technologies.

COMMIT: Factory That Measures Bad Work

COMMIT is the factory that measures bad work killed. BAND: Planner specs, Builder implements with evidence, Verifier rebuilds clean offline and measures mutation kill + reference-model + invariants. Pocketful double-entry ledger demonstrates it.

Guerrero

Anthropic ClaudeChatGPTBand Agentic Mesh+1

IMBOB Architecture Storytellor

Architecture Storyteller 2.0 and Ripple-Agent is a VS Code and IBM Bob IDE extension that maps polyglot codebases, visualizes dependencies, predicts change impact, generates documentation, and creates architecture dossiers.

LuminexSpark

AntigravityAnthropic ClaudeChatGPT+2

DumbDuck

An agent-agnostic AI execution router that uses ML classification to direct tasks to deterministic local tools, lightweight AI models, or the originating agent—reducing unnecessary model calls while tracking execution performance.

DumbDuck

AI/ML API

CI Intelligence Agent

An IBM Bob-powered CI/CD agent that analyzes pipeline failures, recalls past fixes, checks test and coverage readiness, and blocks unsafe deployments for a smarter, memory-driven workflow.

4play

AgentOpsChatGPTAnthropic Claude+1

Origit

Origit records what an AI coding agent read before it changed your code. A SHA-256-hashed provenance record attaches to every commit via Git hooks. One command traces any flagged input to the sessions, commits and files it may have influenced.

Team Origin

IBMOpenAIVercel+2

RoleCall: Catch Permission Holes Before Merge

RoleCall catches access-control bugs in pull requests before merge: it calls every API endpoint as every role on main and the PR, proves each hole with real HTTP evidence, and IBM Bob finds the exact line and writes the fix plus a test.

team-1680

IBMrest apiChatGPT

FixFlow - Find it,Fix it,Prove it

FixFlow is an AI-powered debugging platform using IBM Bob to analyze code, identify root causes, generate fixes, run tests, and verify results. It transforms debugging into a closed-loop workflow: find the bug, fix it, and prove it.

The Uncompiled

OpenGPTsIBMAI/ML API

Day Zero

Day Zero — AI agent that scans any repo, traces pipelines via 3 subagents, and generates a verified interactive architecture blueprint to onboard engineers in under 2 minutes.

NullGap

IBMAnthropic ClaudeAntigravity

ArchPulse Repair with evidence

ArchPulse makes tangled codebases easier to understand and repair. Interactive dependency graphs give developers and IBM Bob a shared view of the problem, while tests and architecture checks verify the fix.

Spanish team

IBMGPT-5Codex

BlastRadius

BlastRadius maps everything a pull request can break, including callers it forgot to update, and hands that map to IBM Bob, which checks team ADRs and proves each risk with regression tests its subagents write and run.

CoffeeAndCode

IBMIBM GraniteIBM watsonx Assistant+3

BlindSpot

Blindspot exams IBM Bob on your codebase, maps where it is confidently wrong in Bob IDE, and writes targeted context to fix it. In a pre-registered test on tinydb, our context raised Bob from 89% to 95% accuracy (p = 0.017).

Boundary

Claude CodeGemini AIIBM

BugDNA

BUG DNA traces today's bug through your codebase, Git history, and past fixes to reveal whether you're about to fix the same bug twice.

GitHavoc

IBMVercelGithub Copilot+1

Bob-Marketplace

Bob Marketplace is the first decentralized package manager and web registry for IBM Bob 2.0, enabling conversational in-IDE discovery, one-click installation, and lifecycle management for AI Skills, Modes, Rules, and MCP Servers.

nodesemesta

IBMVercel

Smart Repository Onboarding Assistant

An AI-powered assistant built with IBM Bob that accepts local folders/ZIPs, offers unlimited code Q&A, auto-generates READMEs, and fixes code instantly.

Smart Repository Assistant

StreamlitGemini AIGemini 3 Flash+2

Adversarial Migration

Two IBM Bob agents translate COBOL to Python while being completely blind. A prosecutor hunts for inputs where they both disagree, and the real COBOL binary judges byte-for-byte. Found 4 silent bugs, then verified 5 programs on 5,044 inputs.

Bay Harbour Butcher

Anthropic Claude

ShiftAble

ShiftAble is an AI-powered accessibility simulator that automatically audits UI source code against WCAG 2.2 standards using Motor, Visual, and Vestibular/Cognitive personas then provides actionable and patch-ready fix reports directly.

labyu

IBMAntigravityChatGPT

AI Fuel

AI Fuel is an AI resource-management platform that helps developers plan, prioritize, and complete tasks within a limited AI token budget. It shows what requests require, what can be completed, and how available AI capacity is being used effectively now.

Tech Duo

ChatGPTClaude CodeGemini 3 pro+6

CodeTuner, your assistant with old projects.

CodeTuner is an IBM Bob IDE extension for measurable AI code optimization. It uses parallel analysis, mandatory approval gates, and automated benchmark regression checks to modernize code and boost performance without breaking builds.

ShipFast

Gemini AIChatGPTIBM

Agentic Local Version Control System (LVC)

An agentic local development workflow assistant, powered by IBM Bob to streamline coding, debugging, and task automation directly in your local environment.

AZIU Kod

Gemini 3 pro

RuleScope: Know Before You Code

A custom IBM Bob mode, skill and MCP server that turns any change request into an evidence-based Impact Brief: every file and line it touches across repositories, hidden blockers, a tested before/after of the naive fix, and a safe plan.

Kinetic

VercelIBMAnthropic Claude

Tursiops

Tursiops is a VS Code plugin adding a memory and navigation to your AI coding agent using your Gemini API key. Has a mini version control, find files via natural language, summarize changes between git commits, and auto saves snapshots to revert changes.

Conquest Core

Generative AI StudioGithub CopilotIBM+3

PromptGold

PromptGold × IBM Bob: prompt QA that runs itself. Bob reads your prompt and its policy, writes the tests, attacks the prompt with parallel subagents, patches the leaks, and opens the PR — while a hook keeps the final "bless" in human hands.

MBG (Massive Burnout Guaranteed)

Claude CodeIBMAI/ML API+3

Version Fault

Version Fault is a VS Code extension that traces software regressions from a broken release to the most likely bad change, then uses IBM Bob to explain the root cause, apply a targeted fix, and verify it with tests.

Hackuna-Matata

IBMAntigravityChatGPT+1

ProofPatch Fix bugs. Show proof.

A reproducible verification workflow for AI-assisted repairs, built for the IBM Bob 2.0 Hackathon. The live app executes trusted bundled Python code; no AI service or API key is required.

ProofPatch

StreamlitLlama 3.1GPT-5+2

Proof Carrying Review

An IBM Bob mode that reviews a pull request against its ticket and spec and reports nothing unless it wrote a test that fails because of it. Every finding arrives with proof, so reviewers stop guessing which AI comments are real.

Two Faced Coin

Anthropic ClaudeClaude CodeCodex+4

Bobtention : A Human Attention Routing

A human attention router for IBM Bob that lets Bob code autonomously and interrupts developers only when detecting task drift, repeated failures, scope creep, or high-impact actions.

Saya Akan Lawan

AIMLIBM

Rowgate: the signed contract as a release gate

Rowgate reads the API contract your partner actually signed - a spreadsheet - cites the exact cell a release branch breaks, and writes the failing test that proves it. Built with IBM Bob. A human decides what ships.

Rowgate

Anthropic ClaudeIBMVercel

Hall Monitor: Code Review of AI-Written Changes

Hall Monitor supervises IBM Bob while it codes: it enforces rules from your policy documents on every action, rejects excuses before they run, and makes Bob back every claim with receipts, so review starts from evidence.

1stdayalive

IBMClaude CodeBright Data MCP Server+1

Intelligent code review coach

Intelligent Code Review Coach analyzes code changes before they are merged, detecting security risks, likely bugs, and weak tests. It runs locally, never executes submitted code, and provides clear findings with practical fixes.

Chile Cut

CodexIBM

AI Production Incident Replay

AI Production Incident Replay is an AI-powered tool that analyzes logs, errors, and recent code changes to reconstruct production incidents, identify probable causes, affected components, and actionable investigation steps.

BOB Minds

Auto-GPT

30apps loaded