hash

HASH is an open-source, multi-tenant platform for building self-building knowledge graphs and simulations. According to its README, it aims to turn raw information into knowledge and process graphs, creating graph-backed "world models" that support process automation, optimization, decision making, and AI. The project describes itself as a self-building, open-source database that grows, structures and checks itself. It integrates data in near-realtime and offers interfaces for browsing and managing entities (data) and types (schemas), including for non-technical users. Autonomous agents can be deployed to grow, check and maintain the database, collating information from the public internet and connected private sources into a standardized semantic form. A notable feature highlighted is bitemporality: first-order support for recording and querying both when something happened and when it became known. The README positions this as central to process modeling. The repository is a public monorepo containing code, docs and resources. Its structure includes /apps (runnable applications and services), /blocks (public Block Protocol blocks), /infra (deployment scripts and infrastructure), /libs (libraries including npm packages and Rust crates), and /tests (end-to-end and integration tests). Getting started options include using the hosted hash.ai service (recommended, quick start under five minutes, with a possible waitlist), running HASH locally, or deploying to the cloud. Local development requires Git, Docker with Compose and Buildx, and mise for pinned toolchains; Docker needs at least 8 GB RAM and roughly 15 GB free disk. Setup involves cloning the repo, installing toolchains and JavaScript dependencies, creating a .env.local with dummy keys (real keys only needed for AI features), starting containerised infrastructure via yarn compose up -d, and then running yarn start or separate graph and dev processes. The graph layer compiles from Rust and can take 10-20 minutes on a cold first build; it can alternatively run in Docker via the hgres profile. Dev mode seeds three users with password "password". The README documents environment variables covering the API server, AWS configuration, file uploads (local filesystem or S3), email transport (dummy, AWS SES, or SMTP), Postgres settings for various services, Redis, optional Rudderstack telemetry, and other options such as frontend URL, session secret, log level, and Block Protocol API key. Most variables have defaults and are not required to run the application. Contributing guidance is provided, along with example IDE configuration files for VSCode and Zed. Most of the repository is published as free, open-source software under licenses detailed in LICENSE.md. Security reporting instructions and contact channels (X, email, discussions, issues) are also listed.