HeadlessOps
Integrations

Knowledge Base

Workspace-scoped Markdown documents for agent and developer context

Knowledge Base

Every workspace has a Knowledge Base — a collection of Markdown documents that describe the workspace's context, conventions, integrations, data sources, and business rules.

AI agents should always consult the Knowledge Base before starting any development task in a workspace.

Label Conventions

Use consistent labels to categorize documents:

LabelUse For
conventionsCoding standards, naming rules, patterns
credentialsWhich credentials exist and what they connect to
data-sourcesExternal APIs, databases, Airtable bases
business-rulesDomain logic, approval flows, SLA rules
architectureSystem design decisions, integration topology
runbooksStep-by-step operational procedures

Managing Documents

List Documents

curl "https://app.headlessops.ai/api/workspaces/ws_abc123/knowledge-docs?search=crm" \
  -H "Cookie: session=YOUR_SESSION"

Returns an array of { id, title, labels, snippet, updatedAt }.

Get a Document

curl https://app.headlessops.ai/api/workspaces/ws_abc123/knowledge-docs/doc_id \
  -H "Cookie: session=YOUR_SESSION"

Returns the full Markdown content.

Create a Document

curl -X POST https://app.headlessops.ai/api/workspaces/ws_abc123/knowledge-docs \
  -H "Cookie: session=YOUR_SESSION" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Airtable CRM Schema",
    "content": "# Airtable CRM\n\nBase ID: appXXX\n\n## Tables\n...",
    "labels": ["data-sources", "crm"]
  }'

Update / Delete

# Update
curl -X PUT https://app.headlessops.ai/api/workspaces/ws_abc123/knowledge-docs/doc_id \
  -H "Cookie: session=YOUR_SESSION" \
  -d '{ "content": "...updated content..." }'

# Delete
curl -X DELETE https://app.headlessops.ai/api/workspaces/ws_abc123/knowledge-docs/doc_id \
  -H "Cookie: session=YOUR_SESSION"

In-Integration Knowledge Files

In addition to workspace-level documents, integrations can include a knowledge_base.md file in their package. This file documents integration-specific context: API schemas, field mappings, business rules, and notes for future developers or agents.

See knowledge_base.md and fix_log.md conventions for the recommended format.