One canonical research store
ThothTerminal centers the work on a canonical SQLite system of record rather than on scattered caches, duplicate databases, or prompt-only memory.
Active system
ThothTerminal
A local-first research terminal for heterogeneous sources, canonical storage, structured analysis, and report-grade outputs. Thoth is built for people who need companies, supply chains, transcripts, and research artifacts to stay coherent as the corpus grows.
Current surface
What ThothTerminal is already doing.
The product has moved well past a generic research dashboard. The repo now points to a deliberately structured terminal with ingestion, triage, transcripts, earnings, report work, and MCP exposure as part of one system.
Web, desktop, MCP, and Discord
The same research system is exposed through the browser, a desktop shell, MCP tools, exports, and optional Discord workflows instead of fragmenting into unrelated interfaces.
Companies, earnings, transcripts, supply chains
The current UI includes company pages, earnings views, transcript lookup, currency normalization, datasets, and supply-chain graph work.
Report-grade exports
The roadmap increasingly treats the research report as the canonical output surface, with versioned reports, knowledge assets, export history, and freshness review.
Build doctrine
Why the architecture matters.
The strongest part of the ThothTerminal materials is not a slogan but the operating doctrine: one durable core, structured state, headless workflows where possible, and research outputs that can be audited instead of merely admired.
Cheap enough to stay on
The design bias is toward a serious always-on terminal with bounded polling, explicit schemas, and low idle cost rather than heavyweight infrastructure.
Shared business logic over forks
Web, desktop, automation, and MCP surfaces are supposed to reuse the same logic and data contracts, so the system does not drift by interface.
More than a dashboard
The long-term target described in the roadmap is an institutional intelligence system that can build and continuously refresh investment-grade views of companies, programs, and bottlenecks.