BURST · EST. 2023
Investment research,
powered by AI and software
Founded in 2023, Burst develops AI-assisted investment analysis tools and related software. Our goal is to bring charts, market context, and research notes into a workflow where people can check the evidence behind their conclusions.
THE PROBLEM WE ARE WORKING ON
From scattered information to research people can review
Market research involves moving between price charts, news, technical indicators, and analysis notes. It can be difficult to keep track of which source supports a conclusion, when information became available, and which assumptions a historical test uses.
Burst's development focuses on connecting those parts: inspect the data, compare a fixed rule, and review the context behind an interpretation.
WORKING LOCAL PROTOTYPE
Charts, research, and historical testing in one workflow
We are developing and testing a research prototype that runs locally. It combines market charts, technical indicators, historical rule comparisons, and timestamped market context.
Charts and technical indicators
Explore candlestick charts across multiple symbols and timeframes, with indicators including moving averages, MACD, GMMA, and OBV.
Historical rule comparisons
Compare fixed rules using completed daily bars and next-session opening prices. Review the test period, transaction cost assumptions, exposure, and drawdowns alongside the results.
Market context
Review market observations with source references and timestamps alongside the charts. The prototype distinguishes information available for forward research from material with a publication timestamp suitable for historical review.
AI-assisted analysis research
We use GPT to explore trading analysis workflows and investigate how AI and software can help organize research for human review.
The prototype is a research tool with no order execution features. Historical tests evaluate past data and do not predict future returns.
OUR APPROACH
Use software for calculations and AI for research context
The charts and historical comparisons are part of the current prototype. AI-assisted extraction, synthesis, and source-linked notes are the next development goals. We are exploring how to connect these steps while keeping the underlying evidence visible.
Current prototype: charts, indicators, and fixed-rule historical comparisons with explicit assumptions.
Planned AI integration: extract observations from source material and draft notes with timestamps and references.
Planned evaluation: compare each generated statement with its source, flag missing support, and review the draft before use.
Planned Claude integration
We plan to use the Claude API to extract structured observations from research material and draft notes that retain source references and timestamps. Software will handle numerical indicators and historical calculations, while people review the supporting evidence and AI-generated notes.
Our next milestones are API integration, evaluation against the original sources and our current GPT-assisted workflow, and refinement of Japanese-language research notes. We will assess factual accuracy, source attribution, temporal consistency, latency, and cost.
About Burst
Business name
Burst
Business type
Sole proprietorship
Founded
2023
Team
5 people
Focus
AI-assisted investment research software
Contact
Business inquiries
Building practical research workflows
Burst develops software for investment research and market analysis. Our current work centers on a local research prototype and AI-assisted analysis workflows, combining visual market data with context and historical testing.
Business and development inquiries
hello@burst-research.net
Burst · Founded in 2023
Contact Burst