documentation

Use PodsHub from direct tools and bridge-friendly harnesses.

PodsHub gives buyers cheaper AI coding usage through standard /v1-compatible endpoints for chat, Codex Responses, and Anthropic-style messages, with bridge paths when a tool still wants a local router.

pick your toolcreate API keysee pricing
endpoint
https://api.podshub.xyz/v1
Auth
Bearer key
Format
OpenAI / Anthropic
start here

Quickstart

01

Create an account

Sign in to PodsHub, then open the dashboard. Your buyer workspace is ready immediately after registration.

create account->
02

Top up balance

Create a Pending Top-Up for operator review before sending traffic.

open wallet->
03

Create a key

Generate a PodsHub key and keep it private. Paste the setup prompt into your harness, then give it this key when asked.

create key->
clients

Tool setup

Tool
Where to configure
What to paste
Hermes
Paste setup prompt
Prompt configures a custom OpenAI-compatible endpoint with PodsHub values.
OpenClaw
Paste setup prompt
Prompt uses the custom API-key provider path with OpenAI compatibility.
OpenCode
Paste setup prompt
Prompt adds a podshub provider using @ai-sdk/openai-compatible.
9router
Paste setup prompt
Prompt adds PodsHub as an upstream, then exposes localhost:20128/v1.
Claude Code
Paste setup prompt
Prompt uses PodsHub's Anthropic-compatible /v1/messages route.
Codex CLI
Paste setup prompt
Prompt configures a Responses provider and explains Codex Speed.

For guided setup, use the tool picker on the Use page. Hermes, OpenClaw, and OpenCode use OpenAI-compatible chat; Claude Code uses `/v1/messages`; Codex CLI uses `/v1/responses`.

reference

API basics

Base URL
https://api.podshub.xyz/v1
Authorization
Bearer ph_live_...
OpenAI chat
POST /chat/completions
Codex CLI
POST /responses
Claude Code
POST /messages
Codex Speed
Fast sends service_tier: priority
Planned discovery
GET /models
Content type
application/json
curl https://api.podshub.xyz/v1/chat/completions \
  -H "Authorization: Bearer ph_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      { "role": "user", "content": "Write a small test plan." }
    ]
  }'

Codex Speed

Standard omits service_tier. Fast is separate from Reasoning and sends service_tier priority when the selected upstream supports it.

{
  "model": "gpt-5.5",
  "input": "Write a tiny smoke test.",
  "service_tier": "priority"
}
buyers

Billing and usage

Buyer balance

Requests spend from your buyer balance. Add funds before running agents so long sessions do not stop mid-task.

Usage records

The dashboard tracks usage so you can inspect spend by key, model, and recent activity.

Local payments

PodsHub records Pending Top-Ups through the blank adapter until a live payment provider is configured.

Pricing

Prices can vary by available seller capacity and model. Check the pricing page before large runs.

marketplace

Seller guide

01

Apply

Seller access is reviewed before capacity is listed.

02

List capacity

Add eligible spare usage and follow the dashboard setup instructions.

03

Earn

PodsHub tracks seller earnings and supports payout requests when ready.

open seller dashboard
support

Troubleshooting

401 Unauthorized
Check that the key is copied completely and sent as an Authorization bearer token.
402 or balance error
Top up buyer balance, then retry after the payment status updates.
404 model not found
Use a model currently available in PodsHub pricing or dashboard model options.
429 or capacity error
Retry later or choose another available model/provider route.
Unexpected client error
Confirm your client is using the /v1 base URL, or use a bridge when the client sends non-OpenAI payloads.

Still blocked? Send the failing request shape, model name, and approximate timestamp through the contact page.