Which API should you use with Octop?

For Octop, Tencent Cloud's self-hosted multi-user assistant, the best API for a household or small team that chats with it every day is a cost-optimized OpenAI-compatible provider: in Model Management → Custom Providers → New Provider, set Type to OpenAI Compatible, Base URL to https://api.stdcmpt.com/v1, add your key and the model standardcompute, and Standard Compute's router sends each step to the cheapest model that can do it (Claude Fable 5.1 or GPT-6 Astra only when it counts), on a plan from $19 a month with a $20 compute budget. Octop's presets (DeepSeek, Zhipu, Kimi, OpenAI, Anthropic, Tencent Cloud plans and more) bill per provider, and Ollama runs local models for free. Octop itself is free and MIT-licensed.

Light or occasional use

If one DeepSeek model handles what your household asks, Octop's DeepSeek preset on a pay-per-token key will likely cost the least. A local model through Octop's Ollama preset costs nothing per token if the machine is strong enough.

Daily use, cost-optimized

When several users and their expert agents, scheduled jobs and chat channels share one Octop, a Standard Compute provider sends each step to the cheapest model that can do it, and for every model we vet providers for price, uptime and speed, with automatic fallback. Opt-in Europe or US Compliance rules apply per key, team or company, and we don't train on your prompts or store them. The bill stays at the monthly plan price: requests stop with HTTP 402 when the compute budget is used, so a busy month cannot bill you on top.

API and subscription options for Octop

Prices and allowances can change. Octop's custom-provider fields, Auto mode and stream-usage default come from its dashboard source (the New Provider dialog and its English strings) and its user guide, read on October 15, 2026. Octop user guide: model providers · Octop CLI reference · Octop repo. Spot an error? Tell us.

Add Standard Compute as an Octop custom provider

Name      standardcompute
Type      OpenAI Compatible
Base URL  https://api.stdcmpt.com/v1
API Key   sc_live_…              (your Standard Compute key)
Models    standardcompute        (or Fetch models, then pick)

Run Test Connection and save, then choose Set as preferred on standardcompute in the model pool. Model Management sits in the admin part of the sidebar. Leave live token usage off for the provider unless you want it: Octop starts custom providers with it off.

Octop questions

Does Octop's Auto mode pick the cheapest model?

No. Octop's Auto mode keeps your preferred model and switches to another enabled model only when a message holds content the preferred one cannot handle, such as an image. To have each step go to the cheapest model that can do it, set a routing provider as preferred: with Standard Compute's model standardcompute the router makes that choice per request (Claude Fable 5.1 or GPT-6 Astra only when it counts).

Why does an Octop chat fail on my custom provider when live token usage is on?

Octop's dashboard notes that some private gateways reject the stream-usage option, which fails the chat, so custom providers start with it off and bundled presets with it on. Standard Compute serves OpenAI-compatible streaming at https://api.stdcmpt.com/v1; if a chat fails right after you turned live usage on for a custom provider, turn it off again and retry.

Can Octop use an Anthropic-compatible endpoint?

Yes. A custom provider's Type can be OpenAI Compatible, Anthropic or Bedrock. For Standard Compute, OpenAI Compatible with https://api.stdcmpt.com/v1 is the simpler choice, and Fetch models only works for that type; Standard Compute also serves Anthropic Messages at https://api.stdcmpt.com.

Can I use Standard Compute with Octop from mainland China?

Claude, GPT and other models are subject to the regional policies of Anthropic, OpenAI and the other model providers, and mainland China is outside their supported regions. Standard Compute is for developers in regions those providers support, for example Chinese-speaking teams in Singapore or Malaysia. Inside mainland China, Octop's DeepSeek, Zhipu, Kimi and Tencent Cloud presets are the practical route.