Understand the project
Practical overview
When it fits
Use it when an AI-assisted research or engineering task needs a persistent evidence record, explicit challenge of assumptions and a verification-oriented next-step process.
Who it is for
- Researchers using AI-assisted workflows
- Engineers working with Claude Code, Codex or other agent hosts
- Developers building evidence-aware agent workflows
- Teams that need research context to persist across model sessions
Documented uses
- Record claims with sources, evidence tiers and provenance
- Detect conflicts, weak support and gaps in an answer
- Carry research context between models and sessions
- Separate agent-continuable work from decisions or actions requiring user input
- Expose evidence and verification workflows through MCP
Before you choose
Requirements & limits
The repository is public and MIT-licensed. The documented quick start requires Node.js 24 or later. v2.0.4 is distributed as a GitHub release rather than an npm registry release.
- The local ledger is designed to remain with a task across models and sessions.
- The workflow records typed claims, sources, evidence tiers and provenance.
- CLI, Claude Code and Codex plugins, and MCP access support different agent-host arrangements.
- The MCP connection requires an MCP-compatible client, and the supplied documentation does not identify the complete tool surface.
- The documented setup uses a GitHub checkout and local Node.js environment.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Evidence and transparency
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 3 October 2026
- Original website title
- grainulation/grainulator · Official homepage title · retained in its source language