Coastal Digital Research
Open-source infrastructure for AI agents. Same content, two shapes: HTML for humans, JSON and Markdown for machines.
Pages
Mission
Why we started Coastal Digital Research and what we're working on.
Partnership
We train and publish models for outside researchers. SGJM and Arboris, both by Adam Pippert, are the first.
Policies
Open source, data, and governance policies.
Projects
Open-source repos we maintain under Coastal Digital Research.
Research
What we're thinking about and writing about.
Workflows
How CDR actually runs.
Agent Registry
Agents we run. Manifests are public.
cdrcache
Content-addressed cache for AI agent outputs. Store and retrieve deterministic agent results by input hash.
python3.12mae
Persistent AI agent for long-running development and operations tasks on Linux systems.
python3.12rlm-linux
Remote Linux machine agent exposing a standardized interface for AI systems to operate on Linux hosts.
python3.12Machine-Readable Endpoints
Every human page has a structured-data twin. Bots can also just send Accept: text/markdown or Accept: application/json to any page URL.
GET /.well-known/agent.json: discovery manifestGET /agent/pages.json: all pages as JSONGET /agent/page/{slug}.json: one page as JSONGET /agent/page/{slug}.md: one page as MarkdownGET /agent/agents.json: all agent manifestsGET /agent/agents/{name}.json: one agent manifest