InFinder structures market intelligence into an interconnected graph of directional patterns, investment theses, and ticker positions — with typed relationships, cascade detection, and full audit trails.
Press ↵ Enter or click Search
Type a query and press Enter to search
Every market pattern, investment thesis, and ticker position lives in a typed knowledge graph. This isn't a dashboard — it's a reasoning engine that exposes hidden dependencies across your entire investment universe.
Not just a visualisation — a computational structure that enables capabilities impossible with spreadsheets or flat databases.
A knowledge graph isn't just a visualisation — it's a computational structure that enables capabilities impossible with spreadsheets or flat databases.
When "Fed Pivots Dovish" strengthens, the graph traces impact across every linked perceptron, scenario, and position — instantly surfacing which theses need revisiting.
The graph explicitly flags when perceptrons contradict each other. If "Consumer Spending Weakening" and "Retail Revenue Growing" both exist, InFinder forces intellectual honesty.
If 80% of your scenarios depend on one perceptron and it weakens, the graph makes hidden danger visible — showing exactly how many theses collapse at once.
Signal strength decays through the graph with ML-learned rates. Recent evidence weighs more — the system optimises decay from actual market outcomes.
InFinder structures the full pipeline — every step automated, every decision auditable, every link traceable back to source content.
InFinder doesn't deal in vague sentiment. It structures market knowledge into perceptrons and scenarios — each scored, linked, and auditable.
A perceptron is a structured, directional statement about a market force — not a keyword or sentiment, but a persistent, observable pattern extracted by AI from content.
Each carries a dual score: Importance (significance for markets) and Probability (likelihood of persisting). The system discovers typed relationships — one may reinforce, contradict, validate, or invalidate another.
A scenario is an investment thesis built from multiple supporting perceptrons — a structured argument for why specific assets should be long or short.
Scenarios are AI-generated or manually curated, tracked with entry prices, and scored against actual outcomes. Every scenario links back to its source perceptrons — creating a full audit trail from content to trade.
Three specialised agents operate directly on the investment graph — augmenting human judgment, never replacing it.
Runs a 10-node agentic analysis workflow per ticker: technical analysis, fundamental screening, news assessment, graph context, and overall synthesis.
Autonomously manages portfolio construction: proposes position changes, links perceptrons to scenarios, and deprecates stale theses — with human-in-the-loop approval.
Detects unexplained ticker movements and autonomously searches for information that could explain them — surfacing new perceptrons the graph doesn't yet contain.
A complete AI pipeline from raw information to investment action — every step automated, every decision auditable.
Continuously monitor YouTube channels, Reddit, RSS feeds, and manual inputs. Automated polling, transcript extraction, and engagement scoring.
AI extracts 3-7 directional market patterns per content item. Dual Importance/Probability scoring, 4-context vector embeddings, automatic linking.
Typed relationships (reinforcing, contradicting, validating, invalidating) across all entities. Multi-hop reasoning and cascade detection.
AI-generated or manual investment theses. Clusters related perceptrons, suggests L/S positions with rationale, and tracks confidence over time.
Every scenario links to source perceptrons, every perceptron to source impulses, every impulse to source content. Compliance-grade traceability.
Learns optimal score decay rates from market outcomes. Predicts scenario performance. Source importance scoring based on perceptron outcome quality.
The only platform that covers the full content-to-investment-thesis pipeline with a typed knowledge graph.
From hedge funds and RIAs to independent research analysts — InFinder augments the investment process at every level.
Reduce 200+ daily sources to scored directional patterns. See concentration risk, cascade exposure, and thesis contradictions at a glance. Full compliance audit trail.
Structure your research into linked perceptrons. Surface counter-evidence to your thesis automatically. Track which patterns persist vs decay over time.
Every investment decision traces back through scenarios → perceptrons → source content. Graph-based concentration analysis. Revertible agent actions with full state logging.
Search across 275878 nodes and 239807 edges — or begin adding your own content sources.