Research & KnowledgeDone with custom-built or unnamed agent Apr 14, 2026
Run persistent financial research with an agent
LangAlpha is a Claude Code-style agent for investing research that keeps persistent workspaces and memory so analysis such as models and comps can be updated across sessions.
- What They Did
- You create one workspace per research goal, such as a Q2 rebalance. The agent interviews you about goals and style, then writes code in a Daytona sandbox to pull financial data and build deliverables like DCF or comps models. It saves files and an agent.md notes file in the workspace. Parallel subagents handle sub-tasks, and you can steer mid-run.
- How It Went
- The README gives no measured results. It shows a comps run that produced an Excel model and PDF report with implied price ranges for NVDA, AMD and GOOGL. Later sessions resume from the saved files and notes.
- Worth Knowing
- You can self-host with just Docker and your own LLM key, but without data keys you get reduced Yahoo Finance data. Price-triggered automations and channel integrations are hosted-only.
Try it yourself with custom-built or unnamed agent
Create a persistent research workspace for [company or ticker] in [folder]. Gather its latest filings and news, build a simple financial model and a comps table against [peers], and write notes on assumptions in a memory file. Next time I ask, update the model with new data and summarize what changed. Finish by showing the first model and notes.
Do this with a skill.
These SkillShop playbooks do this exact job.