Stryda Documentation

What is Stryda?

Stryda sits between your AI agents and the tools they call. Every request runs through a fixed pipeline — validate → policy → approval → adapter → audit — enforced on the backend, not trusted from the client. Agents get deterministic authorization, operators get a tamper-evident record, and compliance teams get evidence on demand.

Prerequisites

Before you start, you'll want the following:

Quick start

Get your first agent under governance in about five minutes.

1. 1

   Create an account

   Sign up at stryda.ai with email + password or Google SSO. You land on a dashboard with sample workflows, chatbots, and agents pre-loaded so the UI is never empty.

2. 2

   Add an AI provider key

   Open Settings → API Keys, pick Anthropic / OpenAI / Google, paste your key, save. At least one key is required before any AI step can execute.

3. 3

   Create a workflow

   Click "+ New Workflow" or pick a template. The visual editor opens with a Trigger node. Drag AI Action, Condition, Output, or Approval nodes from the toolbar.

4. 4

   Configure AI nodes

   Each AI Action node picks a model (e.g. Claude Sonnet 4.6), a prompt, and optionally temperature + max tokens. Reference upstream data with {{node_id.output}}.

5. 5

   Run a test

   Click "Run" — each node lights up as it executes. The Run History panel shows inputs, outputs, token usage, and the exact cost per step.

6. 6

   Publish and integrate

   Publish the workflow, then open the Integrate tab for API, webhook, cURL, JavaScript, and Python snippets to trigger it from any system.

Trigger a workflow run via the REST API

curl -X POST https://stryda.ai/api/v1/workflows/WORKFLOW_ID/run \
  -H "x-api-key: ${STRYDA_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{"input_data": {"prompt": "Summarize this customer email"}}'

const res = await fetch('https://stryda.ai/api/v1/workflows/WORKFLOW_ID/run', {
  method: 'POST',
  headers: {
    'x-api-key': process.env.STRYDA_API_KEY,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    input_data: { prompt: 'Analyze this support ticket' },
  }),
});

const run = await res.json();
console.log(run.run_id, run.status); // "run_abc123" "running"

import os, requests

res = requests.post(
    "https://stryda.ai/api/v1/workflows/WORKFLOW_ID/run",
    headers={"x-api-key": os.environ["STRYDA_API_KEY"]},
    json={"input_data": {"prompt": "Analyze this support ticket"}},
)
run = res.json()
print(run["run_id"], run["status"])  # run_abc123 running

Explore the docs

Key concepts

Supported AI models

Stryda routes prompts to models across three providers. Pick per step based on quality, speed, and cost — every token is logged to the audit ledger with its USD cost.

Troubleshooting

Next steps