AI experiments, agents, and developer tooling.

Portable AI code review workflows you can drop into any repository — production-focused review for bugs, regressions, security, performance, and missing tests.

Codex-style review prompts including codex-code-review (production-focused review for bugs, regressions, security, performance, missing tests) and codex-security-audit (threat-model driven audit for injection, auth, secrets, data exposure, config, dependency risk). Built for practical solo-developer workflows: Android, Kotlin Multiplatform, web apps, CLIs, scripts.

Benchmark AI coding agents head-to-head — paste a task, pick the agents, see who wins. Ranked leaderboard with per-agent scoring across speed/quality/efficiency.

Paste any coding challenge, select the AI agents you want to compare, and get a detailed battle report: ranked leaderboard, per-agent scoring, step-by-step breakdown, and an AI-generated recommendation on which agent to use for what.

Design multi-agent AI systems from a single goal — get a complete agent network blueprint with roles, tools, workflow steps, and implementation code.

Describe what you want your agent system to accomplish, pick a framework, and get a full multi-agent architecture: system name, overview, per-agent cards (role/tools/responsibilities), numbered workflow, and a collapsible implementation code block in your chosen framework.

Turn rough ideas into production-ready prompts — optimized prompt, system prompt, improvement breakdown, variants, test cases, and anti-patterns, all in one forge.

Paste your rough prompt, pick use case/target model/tone/output format, and get a fully engineered result: quality score, optimized prompt, optional system prompt, per-area improvement explanations, technique tags, prompt variants for different scenarios, test cases, and anti-patterns to avoid.

Simulate offline caching behavior for any digital signage playlist — know what fails before it fails on screen.

Paste a playlist configuration, set device bandwidth and storage, pick a failure scenario, and get a full simulation: cache score, offline readiness verdict, step-by-step caching process, risk breakdown per content item, and actionable fixes.

Simulate how your digital signage content appears across different screen setups before you deploy — readability scores, brightness feedback, distance analysis, and a pre-deploy checklist.

Describe content, pick screen count, orientation, brightness, viewing distance, and environment. Get a full simulation report: overall score, per-screen readability and color accuracy, common issues, brightness and distance feedback, orientation assessment, content adjustments, and a pre-deploy checklist.

AI-powered QA tool for digital signage campaign assets — catch issues before content goes live on screen.

Upload a campaign image, set display size, audience, and goal, and get a QA report: overall score, per-check scoring across readability/contrast/sizing/branding/content/technical dimensions, critical issues flagged, and specific recommendations.

AI readability and visibility checker for digital signage — analyze content at real-world screen sizes and viewing distances before you deploy.

Optionally upload a screenshot, pick screen size, viewing distance, and environment, and get a full readability report: readability score, visibility rating, minimum font size recommendation, contrast rating, clutter score, safe zone compliance, per-check results, and improvement suggestions.

See any topic through multiple disciplinary lenses simultaneously — scientific, philosophical, economic, cultural, historical, and ethical.

Type any question or topic, select 2 or more perspectives, and get a structured multi-lens analysis: a confidence-weighted view from each discipline, areas where perspectives agree, points of conflict, and a final synthesis that bridges them.

AI-generated deployment checklists for digital signage installations — tailored to your exact hardware, network, and environment.

Configure player hardware (BrightSign, Samsung SSSP, LG webOS, Raspberry Pi 4, Intel NUC, Chrome OS, Android, Windows), screen count, network, mounting, content, location, CMS, and internet reliability. Get a risk assessment, time estimate, and a three-phase interactive checklist with critical-item badges, tips, and progress tracking.

Replay and understand your dev sessions from terminal history and git logs — turn raw output into structured insight.

Paste terminal history and git log output, hit Replay Session, and get a structured breakdown: chronological replay steps, session phases, errors with fixes and lessons, and key learnings — all extracted by AI.

Check if any URL can be embedded in digital signage — instantly know compatibility, CSP issues, and workarounds.

Paste any URL, choose your check type, and get a full embed compatibility report: overall score, per-check breakdown (X-Frame-Options, CSP, autoplay, HTTPS, auth), platform-specific compatibility notes for major signage CMS platforms, and actionable workarounds with difficulty ratings.

Diagnose why any URL won't embed in an iframe — HTTP header analysis, plain-English explanation, and copy-ready fix snippets.

Paste a URL, choose your digital signage platform, and get a full iframe compatibility diagnosis: overall pass/fail/warning status, per-header analysis with impact ratings, plain-English and technical summaries, an alternative embedding approach, and ready-to-paste Nginx/Apache/JavaScript fix code snippets.

Turn any support thread into a structured escalation report — severity, root cause, timeline, action items, and a ready-to-send customer reply.

Paste any support thread (Slack, email, Intercom, Zendesk) and get an instant structured analysis: incident severity, chronological timeline, root cause explanation, action items by priority (immediate/short-term/long-term), polished customer-facing reply, and detailed engineering notes.

Transform raw support notes into polished customer replies and internal notes — in one click, for digital signage SaaS support teams.

Paste messy technical notes, pick a tone and account context, and get a fully formatted customer-facing reply, concise internal team note, categorization (billing/technical/feature-request/outage/onboarding), urgency level, suggested next steps, and follow-up date recommendation.

Turn any topic into a complete faceless video script — narration, visuals, B-roll, thumbnails, and tags, ready to record.

Paste a topic, pick niche/platform/style, and get a fully structured faceless video production package: hook, section-by-section script with narration and visual direction, thumbnail concepts, SEO tags, YouTube description, CTA, music mood, and monetization potential.

Optimize your digital signage playlist order in seconds — score, fix issues, and predict engagement with AI.

Add playlist items, set venue context, and get an AI-powered optimization report: overall quality score, detected issues (errors, warnings, tips), recommended playback order, actionable recommendations, and engagement prediction — tuned to specific venue type and audience dwell time.

Turn any startup idea into a phased execution plan — MVP features, tech stack, pricing, risks, distribution, and week-one actions, all in one shot.

Describe your idea, set stage/revenue goal/timeframe, and get a full AI-generated startup playbook: product name, one-liner, MVP feature list, recommended tech stack, multi-phase execution roadmap, pricing tiers, risk mitigations, distribution channels, and exact week-one actions.

Calculate the ROI of digital signage in real time — monthly savings, payback period, 3-year return, and an AI-generated sales summary tailored to your industry.

Adjust sliders for current print costs, staff hours, setup investment, and monthly fees, and watch ROI metrics update instantly. Generate a Claude-powered sales headline, customer-ready summary, top benefits list, and industry-specific insight ready to drop into a proposal.

Detect patterns in digital signage incident logs — black screens, offline events, crashes, and reboot loops — with AI root cause analysis and priority actions.

Paste incident logs from any monitoring system, set network size and time range, and get a full pattern analysis: network health score, systemic vs isolated classification, detected failure patterns with root causes, incident type breakdown, time distribution, affected zones, and prioritized remediation steps.

Load test your digital signage network before deploy — simulate 10 to 10,000 screens and get AI-powered bottleneck detection, scaling limits, and fix recommendations.

Set screen count, bandwidth, content type, scenario, and network type, then run the simulation. Get Pass/Warn/Fail result, network score, bandwidth usage (peak/average/per screen), bottlenecks with severity and fixes, preload timing per screen tier, cache efficiency, scaling limits, and specific recommendations.

Paste device logs and get instant AI diagnostics — root cause, error clusters, timeline, customer reply, and engineering notes, all in one click.

Paste or upload any digital signage device log file, select device type, and get a structured analysis: severity rating, error/warning counts, clustered error patterns with occurrence counts, root cause, incident timeline, copy-ready customer reply, detailed engineering notes, and prioritized next steps.

Discover the right MCP servers for your use case — AI-matched install commands, config examples, and capability lists, tailored to your stack.

Describe what you want to do (send emails, query databases, interact with GitHub), pick your stack, and get a curated list of matching MCP servers with install commands, JSON config examples, and capability tags — ready to drop into your Claude or AI agent setup.

Local-first persistent memory for AI coding agents — direct file writes, zero server in capture path, zero silent failures.

Solves the AI-forgets-everything-between-sessions problem with a local-first memory layer: writes session observations directly to files, syncs across projects, and injects relevant context at the next session start so agents arrive knowing what broke last time.

Generate realistic digital signage mockups for any environment — HTML preview, color palette, font suggestions, and content tips, all from a single config.

Pick environment, screen size, orientation, content type, design style, and brand name. Get a live HTML mockup rendered in the browser plus design notes, color palette, font recommendations, and content tips — tailored to your specific deployment context.

Offline reliability certification for digital signage deployments — simulate power failures, network cuts, reboots, and cache corruption before you ship.

Select device type and deployment context, describe content config, choose which resilience tests to run, and get a full certification report: overall pass/fail, certification level (Enterprise/Standard/Basic/None), per-test results with durations, conditions to maintain certification, critical failures, and validity period.

Generate customized digital signage sales proposals in seconds — executive summary, hardware recommendations, pricing, ROI, and implementation timeline, all tailored to the prospect.

Enter client company, industry, pain point, deployment details, budget range, and timeline. Get a complete sales proposal: recommended solution tier (Starter/Professional/Enterprise), itemized hardware list, first-year pricing, ROI payback period, phased implementation plan, and clear next steps.

AI vision observability for signage fleets — upload a device screenshot and get actual display health, not just a ping.

Drop a screenshot from any signage device, add optional device name and location, and get a real health assessment: status (Healthy/Warning/Critical/Offline), health score, per-issue severity breakdown, content state, display quality score, AI summary, and actionable fix recommendations.

AI-powered device health monitoring for signage fleets — input metrics, get failure predictions, maintenance schedules, and recommendations before things break.

Enter device name, type, location, and key metrics (uptime, storage, memory, cache hit rate, offline/reboot frequency). Get a full health assessment: health score, status (Healthy/Warning/Critical/Offline Risk), per-metric analysis, failure prediction with confidence and timeframe, top issues, recommendations, and maintenance schedule.

Benchmark any digital signage player and get AI-explained performance analysis — FPS, memory, startup time, thermal throttling, and regression detection.

Enter player model, OS, and benchmark metrics (FPS, memory %, startup time, app load time, video smoothness, thermal throttling), optionally provide a previous benchmark for comparison, and get a full performance report: overall score, grade (A–F), per-metric analysis, regression causes, bottleneck identification, optimization steps, and suitability ratings.

AI release confidence scoring for digital signage software — input your test results and get a Ship / Ship with caution / Hold / Rollback verdict with a full breakdown.

Enter version, platform, test counts, crash reports, log snippet, and benchmark notes. Get confidence score (0–100), clear verdict, per-category score breakdown (Test Coverage, Crash Rate, Log Health, Platform Coverage, Benchmark Performance), highlights, blockers, warnings, recommendation paragraph, and next steps.

AI-powered regression detection for digital signage releases — catch issues before they reach screens.

Enter current and previous version strings, device type, firmware change flag, release notes, and test results. Get a full regression analysis: risk level (Low/Medium/High/Critical), risk score, per-area regression predictions with likelihood and test recommendations, breaking changes, safe areas, rollout strategy, ordered test plan, rollback triggers, and executive summary.

Deep technical lens for any repository — paste a URL, README, or package.json and get architecture analysis, health score, risk assessment, code quality scores, and quick wins.

Input a repo URL and/or paste README and package.json content, optionally specify the tech stack, and get a comprehensive technical report: project type, architecture pattern, complexity rating, health score, strengths, weaknesses, risk cards with recommendations, dependency stats, code quality breakdown, scalability assessment, onboarding difficulty, tech debt estimate, and quick wins.

Laptop-only trend discovery and content drafting tool — monitors public Reddit discussions, clusters topics into signals, and helps you write original blog posts using Claude Code.

Zero recurring cost trend monitor. Watches public Reddit discussions, clusters topics into signals, and helps draft original blog posts with Claude Code assistance. Safety guarantees: no auto-posting, no copying full Reddit posts, no publishing without manual approval.

AI-powered visual diff for digital signage — upload before and after screenshots and get a Pass/Review needed/Fail verdict with change-by-change regression analysis.

Upload before/after screen captures, add optional context, and get a full visual comparison: overall diff rating, diff score, per-change cards (area, type, severity, regression flag), regression list, improvement list, verdict, and AI summary. (See also RustScreenDiffAI for the deterministic CLI variant.)

Generate a personalized skill development learning path — milestones, daily plan, curated resources, checkpoints, and career applications, all from a single skill input.

Enter a skill, current level, target level, available time per day, and learning style. Get a complete learning roadmap: estimated total time, phased milestones with topics and projects, day-by-day plan, key resources (books/courses/videos/practice), common mistakes, checkpoints, and career applications.

AI brand compliance reviewer for digital signage content — upload creative, get an approval verdict, brand score, per-category checks, and accessibility rating.

Upload content image, enter brand name/colors/content type/guidelines, and get a full compliance report: approval status (Approved/Approved with notes/Needs revision/Rejected), brand score, per-category checks across Logo/Typography/Color/Layout/Media/Compliance — each with pass/fail/warning and severity, critical issues, suggestions, accessibility score, and urgency.

AI screen placement advisor for any venue — upload a photo or floor plan and get optimal screen locations, mounting specs, and a full deployment strategy.

Optionally upload venue photo or floor plan, pick venue type/primary goal/foot traffic pattern, and get a complete placement report: prioritized screen placement cards (location, screen size, orientation, mounting height, viewing angle, rationale, risks), traffic flow insights, priority zones, zones to avoid, and overall deployment strategy.

AI zone layout analyzer for digital signage — upload a screen layout and get per-zone scores, violations, best practices, and content recommendations.

Upload a screen layout screenshot, specify screen size/orientation/zone count (or auto-detect), and get a full zone analysis: overall layout score, detected zones with individual scores and suggestions, layout type, best practices, violations, accessibility notes, and content recommendations for primary, secondary, and ticker zones.

AI review panel that simulates expert roles to pressure-test plans, PRs, incidents, releases, and product decisions.

Role-based reviewers surface risks, assumptions, missing information, and disagreement before implementation — a safe alternative to cloning real coworkers.

Local-first analyzer for how you use AI coding assistants — surfaces patterns in your developer workflow.

Runs entirely on your machine to study AI coding-assistant usage and reflect back habits, blind spots, and workflow patterns.

AI project workspace with agents, tasks, docs, memory, and tools in one operating layer.

A workspace that brings agents, task tracking, documentation, persistent memory, and tooling together for AI-assisted project work.

AI planning engine that converts PRDs, feature requests, and GitHub issues into tasks, subtasks, dependencies, and execution order.

Developer-focused planning product that turns requirements and issues into an ordered, dependency-aware execution plan.

Build AI agents from natural language — generates architecture, tools, MCP servers, memory design, prompts, and deployment files.

Describe an automation or agent goal in plain language and receive the technical blueprint and scaffolding needed to implement it.

AI project auditor that turns a GitHub repository into security findings, tech debt, documentation gaps, missing tests, and improvement tasks.

Repository understanding and action planning — like an AI-native code-quality scanner combined with a project planner.

AI operating system for projects — converts ideas, PRDs, or repositories into tasks, docs, roadmaps, architecture, and progress reports.

The broadest of the Flow products: a project operating layer that can absorb planning, repo audit, and agent-build workflows.

Last-30-days research ranked by authenticity, not raw engagement — scores every item on signal over hype.

Ranks recent research and writing by authenticity signals rather than engagement metrics, so genuinely useful work surfaces first.

Lossless context compression for LLM agents — fold it down, fold it back.

Compresses agent context without losing information so long-running LLM sessions stay within budget and can be restored intact.

Bonsai mode for AI agents — forces the smallest solution that works. YAGNI, enforced.

A discipline layer for AI agents that constrains output toward the minimal working solution instead of over-building.

Local, privacy-safe observability for terminal-based AI agents, with a status dashboard.

Captures and visualizes terminal-based AI agent activity locally, giving a status dashboard without sending data off the machine.

A local RTS-style dashboard for your AI coding automata — visualizes agent sessions as workers across a living city.

Runs on your machine, starts a localhost dashboard, and renders AI coding sessions as automata working across a living city. First slice includes demo mode, real-time WebSocket snapshots, a Fastify server, and a React/PixiJS dashboard.