OpenAI AI Agent Integration

Overview

ClawStaff uses a Bring Your Own Key (BYOK) model for AI providers. Instead of routing your traffic through a middleman with per-token markup, you connect your own OpenAI API key directly. Your Claws call the OpenAI API at your negotiated rate, and ClawStaff never touches the billing relationship. You get full control over model selection, spending limits, and usage visibility, with zero markup from ClawStaff.

This integration supports GPT-4o, GPT-4, GPT-4 Turbo, and any other model available through the OpenAI Chat Completions API. You can assign different models to different Claws based on the task complexity and cost profile that makes sense for each use case.

What Your Claw Can Do

• Choose the right model per Claw. Assign GPT-4o to Claws that need speed and multimodal capabilities, GPT-4 to Claws that need maximum reasoning, or GPT-4 Turbo for a balance of both. Change the model at any time without reconfiguring the Claw.
• Track token usage. The ClawStaff dashboard shows token consumption per Claw, per task, and per time period. See exactly where your API spend is going and identify which workflows consume the most tokens.
• Set rate limits. Configure per-Claw rate limits to prevent runaway costs. ClawStaff enforces token-per-minute and request-per-minute caps that stop a misconfigured Claw from burning through your budget.
• Use function calling and structured outputs. Claws can use OpenAI’s function calling and JSON mode features to produce structured, machine-readable output that feeds into other integrations and workflows.
• Use multimodal capabilities. With GPT-4o, Claws can process images alongside text, which is useful for analyzing screenshots, reading diagrams, or processing visual content from other integrations.

Who Can Reach Your Claw

The OpenAI integration powers the AI model behind your Claw. Access control is determined by the Claw’s scoping level and the whitelisting on its communication channels (Slack, Teams, Gmail, etc.), not by the model integration itself. However, the BYOK model provides its own layer of isolation:

• By API key. Each API key is isolated to your OpenAI organization. Your Claw’s API calls are billed to your account and invisible to other ClawStaff customers.
• By project-scoped key. OpenAI’s project-scoped keys let you further restrict which project and rate limits apply to each Claw.

The three Claw scoping levels (Private, Team, Organization-wide) are enforced on the communication channels (Slack, Teams, Gmail) not on the model integration. A Private Claw using GPT-4o is private because only you can message it, not because of the API key. The key isolation ensures that your model usage, prompts, and completions are never shared across tenants.

How It Works

1. Generate an API key. In your OpenAI platform dashboard, create a new API key. You can use project-scoped keys to further restrict access.
2. Add the key to ClawStaff. Paste your OpenAI API key into the integration card on the ClawStaff dashboard. The key is immediately encrypted and stored.
3. Assign a model to each Claw. When configuring a Claw, select which OpenAI model it should use. You can change this at any time.
4. Claw calls OpenAI directly. At runtime, the Claw uses your API key to call OpenAI’s API from within its ClawCage container. Requests go directly from the container to OpenAI’s servers. ClawStaff does not proxy, log, or intercept API traffic.
5. Monitor usage. Token consumption and request counts are tracked per Claw and displayed in the ClawStaff dashboard. You can also monitor usage in your OpenAI platform dashboard.

Security

• API key encrypted at rest. Your OpenAI key is stored using AES-256 encryption in ClawStaff’s database. The plaintext key exists only in memory, inside the ClawCage, during task execution.
• Key never logged. ClawStaff does not log, cache, or transmit your API key outside the encrypted store and the runtime container. There are no intermediate proxies.
• ClawCage isolation. Each Claw runs in a dedicated Docker container. Your API key is injected as an environment variable at runtime and is inaccessible to other Claws, containers, or the host system.
• Direct API calls. Requests go from the ClawCage directly to OpenAI’s API. ClawStaff does not act as a middleman, so your key is never exposed to a third-party relay.
• Revoke at any time. Delete the key from OpenAI’s dashboard to instantly cut off access. You can also remove or rotate the key in ClawStaff without affecting your Claw configurations.
• No data retention. ClawStaff does not store prompts, completions, or any data sent to or received from OpenAI. Only aggregate token counts are recorded for billing visibility.

Cross-Integration Workflows

The OpenAI model integration powers every Claw that uses it, enabling all cross-tool workflows:

• Intelligent triage. A Claw powered by GPT-4o can read a Slack message, determine whether it is a bug report or feature request, create the appropriate GitHub issue or Jira ticket, and notify the right team, all using the reasoning capabilities of the underlying model.
• Document generation. When paired with Notion or Confluence, a GPT-4-powered Claw can produce detailed technical documentation, meeting summaries, or project proposals with well-structured prose.
• Code review with context. A Claw using GPT-4 can review GitHub pull requests with deep understanding of code logic, suggest improvements, identify security concerns, and explain its reasoning in review comments.
• Multimodal workflows. A GPT-4o-powered Claw can process screenshots attached to Slack messages or GitHub issues, analyze UI bugs visually, and file detailed reports with both visual and textual context.
• Model flexibility across your fleet. Use GPT-4o for high-volume, quick-response Claws (Slack bots, triage agents) and GPT-4 for deep-reasoning Claws (code review, architecture analysis). Optimize cost and quality across your entire AI workforce.