Scenaryo · Labs – Munich / Remote

    What we recommend, we also build ourselves.

    In Scenaryo Labs we build our own tools and automations – and run them in production. They come out of real consulting questions, at the boundary between consulting reality and technical feasibility. What we design for clients, we test on ourselves first.

    Talk to LabsSee current projectsScenaryo · Munich / Remote
    Many noisy signals resolve into one clear line
    3
    Live projects
    500+
    Sources ingested
    100%
    Autonomous execution
    0
    Middleware layers
    01What we build

    Tools from real consulting questions.

    Labs projects emerge from concrete questions in our day-to-day consulting work. When a problem keeps coming back, a tool emerges – and we publish it once it holds up under load.

    Signalwaves helps you see beyond your own bubble. It gathers the same story from 500+ sources across every political camp, source tier, and world region – and shows you, side by side, how differently they frame it. It makes visible where mainstream and independent, wire and community, East and West diverge. The goal is to open up information camps and build genuine empathy for how other people see the world. We increasingly live in separate informational realities. Each camp reads its own outlets, feels its own emotions about the same events, and rarely sees the other side except as a caricature. At its core, this is a question of empathy. Signalwaves treats disagreement as a valuable signal. For every story it surfaces which source camps clash and which agree, what emotional register each brings, and which regions of the world are even talking about it. By making the spread of viewpoints tangible – and readable in your own language – it turns "the other side" from an abstraction into a set of real, sourced perspectives. Understanding why someone concerned reads the same headline as someone hopeful is the first step toward empathy.

    Download Signalwaves on the App StoreGet Signalwaves on Google Play
    How it works – architecture
    500+
    Sources
    12
    Regions
    hourly
    Refresh
    01

    Global, multi-camp ingestion

    500+ sources – RSS feeds, community and API streams – across six source tiers (Official, Wire, Editorial, Platform, Aggregator, Community), spanning 12 world regions and non-Western perspectives, so no single camp dominates the pool. Around 8,000 articles flow in every hour.

    02

    Stories, rebuilt every hour

    Every hour, multilingual sentence embeddings cluster the last three days of coverage into stories – events reported by at least two independent outlets. A story keeps its identity while it develops and ends after seven days, so the picture shows today's news, not months-old topics. Stories are then sorted into 60+ thematic lenses.

    03

    Emotion & divergence analysis

    Every article is classified by reader emotion (concerning · hopeful · curious · outrageous), by where its outlet is published, by source tier and by language. A divergence metric scores how far source camps pull apart – labelling each story Consensus, Mixed, or Contested.

    04

    Comparison-first interface

    A viewpoint matrix lays opposing camps side by side; the Briefing leads with what's most contested today and how today's mood compares with a normal day; Perspectives holds every story; Newsiverse shows the unfiltered firehose; Lenses let you follow themes. The interface speaks 14 languages, and headlines surface in your language first, English as fallback. Available on the web and as iOS and Android apps.

    Next.jsReactPythonsentence-transformersSQLitePWACapacitor
    Open Signalwaves
    Live · Signal Trading

    S&P 500 Additions Monitor

    Automated signal trading on index rebalancing. When a stock is newly added to the S&P 500, passive index funds are forced to buy it – a predictable demand spike that regularly leads to short-term price increases. Fully automated: RSS monitoring, ticker extraction and limit orders via the Alpaca Trading API.

    PythonAlpaca APITelegram BotRSS
    Live · S&P Index Inclusion Forecast

    Index-Radar

    Index-Radar estimates, for every US stock, the probability of being added to the S&P 500, S&P 100, MidCap 400 or SmallCap 600 within the next twelve months. It screens the entire US equity universe – 6,504 stocks – against S&P's official eligibility criteria (domicile, exchange, free float, liquidity, GAAP earnings, size band) and weights the remaining candidates on nine factors. The probabilities are calibrated so that, per index, they add up to the number of additions measured from the change history. Fully automated: recalculated twice per trading day from free public sources – S&P Dow Jones Indices announcements, Wikipedia, Yahoo Finance.

    Statistical estimates, not investment advice.

    PythonpandasyfinanceGitHub ActionsVanilla JS
    In internal use · Capital Flow Research

    Capital Flow Dashboard

    Every week, the dashboard tracks where capital flows in and where it pulls out – across five asset areas: US tech/AI, broad index ETFs, Europe/defence, gold/crypto/commodities and China. It connects fast fund flows and fund-manager positioning with slow structural flows: AI bond issuance, credit default swap spreads, share buybacks, money market funds and central-bank gold purchases. It shows where price, flow and positioning diverge and states testable triggers; each week it records whether a trigger has occurred, been refuted or is still open. It deliberately gives no buy or sell recommendations. Refreshed every Monday by a scheduled AI research run that reads primary sources such as ICI, the World Gold Council and the Fed H.15 release – 20+ cited sources per run.

    AI researchVanilla JSSVGPrimary sources
    02Why this matters

    Why we build ourselves.

    “We recommend what we use ourselves.”
    — Labs principle, Scenaryo

    Labs keeps us technically close to the reality of our clients. AI transformation works best when you know what is truly feasible today.

    Every project starts with a real consulting question. We build and test it internally and only publish it once it holds up.

    03What's next

    Labs grows continuously.

    Planned: more tools around AI-assisted product development, automated assessments, and data-driven decision support for product teams.

    01

    AI-assisted product development

    Tools for the structured integration of AI into existing product development processes – from use-case screening to implementation support.

    02

    Automated maturity assessments

    Data-based maturity evaluations for product teams, with an automatically generated 90-day action plan.

    03

    Data-driven decision support

    Dashboards and alerting systems for strategic product decisions – from signal capture to structured action recommendations.

    For AI agents

    For AI agents

    We run a public MCP server. AI assistants can use it to query our services, assessments and insights and – only with your explicit consent – submit an enquiry to us. No data is sent without your consent.

    Or let your AI assistant do it

    Paste this prompt into ChatGPT, Claude, or any AI agent. It will talk to our MCP server and file your request directly — a small live demo of what we build.

    Help me contact Scenaryo, an AI consultancy in Munich. Connect to their public MCP server (Model Context Protocol, streamable-http, no auth): https://ai.scenaryo.de/mcp — if you can't attach MCP servers, call it directly via JSON-RPC 2.0 POST requests (methods "tools/list" and "tools/call"). Use its tools to answer any questions I have about their services first. Then collect from me: name, business email, company (optional), topic, and a short message. Ask for my explicit consent to be contacted, and only then submit everything via the "submit_lead" tool — set referring_agent to your platform name and include a 1–2 sentence conversation_summary. Confirm the submission back to me.
    04Get in touch

    Interested in a Lab tool for your company?

    We enjoy briefings at the intersection of consulting question and engineering challenge. Just write to us.

    AI agent? Query us via MCP