plynf

Plynf — agent context optimization layer. Shape large CRM/ERP/Slack tool responses before they hit your LLM.

Package Information

Released: 4 mo ago
Downloads: 2 weekly / 143 monthly
Latest Version: 0.1.0
Author: The Plynf Authors

Available Nodes

Documentation generating…

Plynf
Fetch a tool response (Salesforce, Slack, Order DB, …) through Plynf, which trims it to the fields your agent actually needs.

Documentation

n8n-nodes-plynf

n8n community node for Plynf — the agent context
optimization layer.

Drop the Plynf node into any n8n workflow that calls Salesforce, Slack,
your order DB, or any registered Plynf connector. The node fetches the
data, applies your Plynf shaping policy, and returns a small JSON
payload your LLM step can use directly.

Install

In n8n:

  1. Settings → Community Nodes → Install
  2. Enter n8n-nodes-plynf
  3. Save

Configure credentials

Add a new Plynf API credential:

Field Value
Plynf URL https://app.plynf.com (or your self-hosted proxy URL)
API Key issued at app.plynf.com — free tier available

Use

  1. Add the Plynf node to a workflow
  2. Pick a tool (e.g. Order DB · Get Order)
  3. Provide arguments as JSON (e.g. {"order_id": "12345"})
  4. Pipe the output's result field into your LLM node

The node also returns cache_hit and savings so you can show the
real-time token reduction in the workflow.

Build

npm install
npm run build

Publish

# One-time
npm login

# Publish to npm (already-reserved name n8n-nodes-plynf)
npm publish --access public

Once published, anyone can install it via the n8n UI:
Settings → Community Nodes → Install → n8n-nodes-plynf. No n8n team
approval is needed for npm-installable community nodes. For the
verified-list halo, PR your repo at
n8n-io/n8n-nodes-community.

Workflow templates

templates/customer-support-agent.json — Slack-mention → Plynf-shape →
gpt-4o reply → post back. Import via n8n's
Workflows → Import from File. Replace
REPLACE_WITH_YOUR_PLYNF_CREDENTIAL_ID with the credential id you create
in the n8n credential manager.

Tier behaviour (what happens when you hit Free-tier limits)

The proxy enforces tier gating; the node passes through whatever it gets
back. Free tier caps at 100 000 shaped tokens / month + 3 connectors;
when exceeded the proxy returns HTTP 402 and the n8n step fails with the
proxy's upgrade_hint text. Toggle Continue On Fail on the Plynf node
if you want the workflow to keep going (e.g. fall back to raw data).

License

Apache-2.0

Discussion