Developer Tools

Deploypad

by Geodd LLC

Overview

Connect your Geodd account to view your profile, available AI models, masked API keys, inference usage, and rate limits. Monitor request volume, success rates, errors, average and peak RPM, and end-to-end latency including p50 and p95. View hourly trends over the daily window or daily trends over weekly and monthly windows, across your account or for one API key. Latency percentiles summarize the selected window; trend buckets report mean latency. Inspect GPU capacity and instances, preview GPU launch costs without provisioning, list SSH keys, and add SSH keys. Browse support categories and tickets, create tickets, reply, and close tickets. Prepare inference API keys for explicit browser confirmation in Geodd DeployPad; full key secrets remain in the console and are never returned through chat. Billing changes, payments, and direct GPU provisioning are not supported.

Tools

add_ssh_key

ChatGPT

Save an SSH public key locally, deduplicated per owner and fingerprint. Names must contain only ASCII letters and digits, with no whitespace. Local-only keys may be used by APAC provisioning but are NOT MC-ready: its console requires a synchronized mcKeyName. This tool does not synchronize providers.

close_support_ticket

ChatGPT

Close your ticket. Closing an already closed ticket is a no-op.

create_api_key_handoff

ChatGPT

Prepare a new API key creation request using model IDs returned by list_models. Returns a secure console URL where you can confirm and retrieve the key. The key secret is never exposed through this tool.

create_support_ticket

ChatGPT

Create a ticket. Reuse request_id only for retries of the same request.

get_auto_top_up

ChatGPT

Read your auto top-up status, configuration, and monthly spend.

get_billing_summary

ChatGPT

Read your wallet balance, recent transactions, and daily spending breakdown.

get_limits

ChatGPT

List your current rate limits and capacity configuration.

get_request_usage

ChatGPT

Get your account's recorded inference request outcomes, including failures, RPM and end-to-end latency in milliseconds (median = p50, plus p90/p95). Uses the console analytics: daily (default) has 24 UTC hourly buckets; weekly/monthly have 7/30 UTC daily buckets, including the current partial bucket. Summary percentiles are approximate across the entire window; series latency is a mean. Average RPM includes idle time; peak RPM is the busiest minute. May be cached for 30 seconds. Not token/billing usage or raw request logs.

get_support_ticket

ChatGPT

Read your unexpired ticket and a bounded page of replies.

get_usage

ChatGPT

Get inference usage statistics for the last 30 days, grouped by API key and model.

list_api_keys

ChatGPT

List your API keys with masked secrets and allowed models. Secrets are never returned.

list_gpu_availability

ChatGPT

List GPU types, pricing, and regional capacity.

list_gpu_instances

ChatGPT

List your GPU and baremetal instances.

list_models

ChatGPT

List AI models available to your account, including pricing and capabilities.

list_ssh_keys

ChatGPT

List a bounded page of your SSH public key fingerprints. Private keys are never returned.

list_support_categories

ChatGPT

List available support ticket categories.

list_support_tickets

ChatGPT

List your unexpired support tickets.

preview_gpu_launch

ChatGPT

Create a non-binding cost quote only. Never charges, tops up, reserves capacity, or provisions. MCP launch is disabled; this token is not human consent.