humanhours

n8n community node for HumanHours. Track AI agent tasks and enrich company domains into a labour-cost business case.

Package Information

Released: 4 mo ago
Downloads: 41 weekly / 158 monthly
Latest Version: 0.3.1
Author: Ralf Klein

Available Nodes

HumanHours icon
HumanHours
Track AI agent tasks and enrich company domains into a labour-cost business case

Documentation

n8n-nodes-humanhours

n8n community node for humanhours. Track AI agent tasks straight from your n8n workflows; humanhours computes hours and money saved and exposes them through a dashboard and Reports API.

n8n is a fair-code licensed workflow automation platform.

Installation

Once verified, install the node from your n8n instance under Settings, Community Nodes with the package name n8n-nodes-humanhours.

For local development, follow the installation guide for community nodes.

Credential

Create a Humanhours API credential with:

Field Required Notes
API Key yes hh_live_..., generated at /api-keys in your humanhours dashboard
Base URL no Defaults to https://humanhours.dev. Override only when running humanhours self-hosted or against a local dev instance.

The credential's Test connection button calls GET /api/v1/agents to verify the key.

Operations

Pick a Resource first: Event to track agent runs, or Company to enrich domains for outreach.

Track Event

Send a single agent task event to humanhours.

Field Required Description
Agent ID yes Slug identifying the agent that performed the task, e.g. support-classifier
Task Type yes One of the 50 humanhours built-in keys, or one of your custom keys
Outcome yes Success, Failure, or Needs Review
Agent Duration (seconds) no How long the agent took. Omit for zero.
Human Baseline (minutes) no Override the baseline for this event. Leave empty to use the built-in or org-set baseline.
Model no Model used, ideally the OpenRouter id (e.g. anthropic/claude-opus-4.8). With token counts, humanhours auto-prices the run into net_saved.
Tokens In / Tokens Out no Prompt and completion token counts for the run. Used with Model to compute the real cost.
Metadata no Free-form JSON object attached to the event. Useful for client, ticket_id, model, or token counts.

The node sends an Idempotency-Key header derived from the n8n execution id and item index, so re-runs of the same workflow item are safe.

Company enrichment

Under the Company resource, turn a domain into an outside-in labour-cost and automation business case, then pull the library for outreach.

Operation Calls Notes
Enrich Company POST /api/v1/companies Synchronous (15 to 30 seconds). Fields: Domain, plus Refresh, External ID, Tags. Charges one lookup when the domain is new or Refresh is on.
Get Company GET /api/v1/companies/{domain} Reads one record from your library. Free.
Refresh Company POST /api/v1/companies/{domain}/refresh Re-enriches an owned company. Charges one lookup.
List Companies GET /api/v1/companies Return Format JSON emits one item per company; CSV emits a single csv field. Filter by Tag, cap with Limit. Free.
Queue Bulk Enrichment POST /api/v1/companies/bulk Async. Domains (one per line, max 1000) plus Tags. Returns a job_id; lookups are charged per domain as the worker processes them.
Get Job Status GET /api/v1/jobs/{id} Poll a bulk job until status is done.

Lookups are a hard cap with no overages (Pro 100 per month, Agency 500 pooled). Hitting the limit returns lookup_quota_exceeded until you upgrade.

Usage patterns

Pattern 1, end of a single-node agent. Drop the humanhours node at the end of an OpenAI / Claude / HTTP Request flow. Wire Agent ID to a workflow constant, Task Type to whatever the workflow does.

Pattern 2, branch on outcome. Use an If node to determine Outcome based on the agent's structured output, then merge both branches into a single humanhours node.

Pattern 3, include token cost. Wire Metadata to an expression like:

={{ {model: "claude-opus-4-7", tokens_in: $json.usage.prompt_tokens, tokens_out: $json.usage.completion_tokens} }}

so the dashboard can break down by model later.

Compatibility

  • Tested against n8n 1.69.0 and newer.
  • Requires n8n-workflow (provided by the n8n host).
  • No runtime dependencies.

Resources

Version history

0.3.0

  • Track Event gained Model, Tokens In, and Tokens Out. Pass these (without an explicit Agent Cost) and humanhours auto-prices the run cost into net_saved for real net ROI.

0.2.0

  • Added a Company resource with six enrichment operations: Enrich Company, Get Company, Refresh Company, List Companies, Queue Bulk Enrichment, and Get Job Status.
  • Grouped operations under a Resource selector. Existing Track Event workflows are unaffected.

0.1.0

  • Initial release.
  • Track Event operation against POST /api/v1/track.
  • Humanhours API credential with API Key + Base URL, plus a credential test that calls GET /api/v1/agents.

License

MIT

Discussion