Based — The Foundation

What Based Does

One system that knows everything about your operation

Based started as a session manager and web explorer. It grew into a project registry, then an infrastructure map, then a health monitor. Now it's a self-describing intelligence layer that agents, dashboards, and humans all query the same way.

📋

Project Registry

Drop a .project.yaml in any directory. Based discovers it, indexes its git status, config, sessions, memory files, and artifacts. Convention over configuration — zero setup, automatic discovery.

💬

Session Capture

Every agent session is captured automatically — message counts, tools used, duration, full transcripts. Browse them in the web explorer or query them through the API. Never lose context again.

🖥️

Infrastructure Mapping

Declare your hosts in YAML. Based knows every server, VM, IP, hardware spec, and role in your fleet. Your infrastructure as queryable data — not a spreadsheet you forgot to update.

🌐

Web Explorer

Browse projects, read session transcripts, check git status — all in a web interface. Based isn't just an API, it's a tool you actually open and use daily.

🔍

Cross-Domain Search

One query searches files, configs, sessions, git commits, logs, memory, and infrastructure simultaneously. Results ranked by relevance, grouped by domain, returned in milliseconds.

🤖

Agent-Ready API

A self-describing manifest tells any agent what's available — every domain, field, filter, and example query. Agents discover the API at runtime without documentation. Plus a native MCP server for direct tool access.

❤️

Health Monitoring

Tracks cron jobs, success rates, run history. Know what's healthy and what's failing across your entire operation without checking each system individually.

🧠

Memory & Handoffs

Project memory files and handoff notes are indexed and searchable. Knowledge persists between sessions — agents pick up where they left off with full context.

🔀

Git Intelligence

Tracks branches, uncommitted changes, and recent commits across every project. See which repos need attention without opening a terminal.

12 Domains

Everything in one place

Each domain is a self-contained data source with its own schema, filters, and natural-language example queries. The API describes itself — agents can discover it at runtime without documentation.

📁

projects

Registered projects with git status, tags, and configuration

💬

sessions

Agent session history, messages, and transcripts

🧩

artifacts

Agents, skills, rules, commands, hooks, and MCP servers

🖥️

hosts

Infrastructure hosts, servers, and virtual machines

🔀

git

Repository status, branches, and recent commits

📋

logs

Session logs with level detection and filtering

🧠

memory

Project memory files and handoff notes

⚙️

config

Parsed project configuration across all projects

❤️

health

Cron job health, success rates, and run history

👁️

views

Aggregated artifact views across projects

📊

usage

Activity statistics and operational metrics

🔍

files

Full-text search across all project files

Agent API

Self-describing, agent-native, instantly queryable

The Agent API is the newest layer — a self-describing interface that lets any agent or tool query everything Based knows. Read the manifest once, query anything. No integration code, no custom adapters.

Also available as a native MCP server — agents on the same machine get direct tool access to all 12 domains without HTTP overhead.

GET /api/agent/
Self-describing manifest

GET /api/agent/{domain}
Query a specific domain

GET /api/agent/query?q=
Cross-domain search

GET /api/agent/projects/{name}/full
Full project aggregate

# authenticate
TOKEN=$(curl -s -X POST /api/auth/login \
  -d '{"pin":"..."}' | jq -r '.token')

# search across all domains
curl "/api/agent/query?q=deploy" \
  -H "Authorization: Bearer $TOKEN"

# response
{
  "meta": {
    "total": 45,
    "duration_ms": 312,
    "byDomain": {
      "files": 33, "logs": 4,
      "artifacts": 3, "config": 2,
      "git": 2, "memory": 1
    }
  }
}

How It Works

From filesystem to intelligence layer

Based scans your project directories, captures session data, reads config files, tracks git status, monitors health — and exposes everything through a unified interface.

Drop a config file

Add a .project.yaml to any directory. Declare your project name, deploy scripts, agent settings, and hooks. Based discovers it automatically — no registration, no database.

Based indexes everything

Git repos, session transcripts, memory files, artifacts, health checks, infrastructure — all scanned, parsed, and made searchable across 12 structured domains.

Query from anywhere

Web explorer for humans. REST API for automation. MCP server for agents. Self-describing manifest for discovery. Same data, every interface.