Which API should you use with jcode?

For jcode, the Rust coding-agent harness built for many sessions and swarms, the best API for parallel agents every day is a cost-optimized OpenAI-compatible profile: run jcode provider add standardcompute --base-url https://api.stdcmpt.com/v1 --model standardcompute --api-key-stdin --set-default, 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. jcode also signs in with Claude and ChatGPT subscriptions, which come with usage windows, and with OpenRouter or provider keys billed per token. jcode itself is free and MIT-licensed.

Light or occasional use

If you already pay for Claude Pro or Max or a ChatGPT plan and one session at a time fits inside its windows, jcode's subscription sign-in adds no new bill. For occasional use, a pay-per-token key on an efficient model will likely cost the least.

Daily use, cost-optimized

When you run several jcode sessions or a swarm of workers at once, a Standard Compute profile 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. All sessions share one monthly budget at the plan price: requests stop with HTTP 402 when it is used, with nothing billed on top.

API and subscription options for jcode

Prices and allowances can change. jcode's provider add flags, profile storage and auth-test come from its OAUTH.md and CLI source, and the swarm effort settings from its README, read on October 15, 2026. jcode: OpenAI-compatible API-key providers · jcode docs: providers · jcode repo. Spot an error? Tell us.

Add Standard Compute as a jcode provider profile

printf '%s' "$STANDARDCOMPUTE_API_KEY" | jcode provider add standardcompute \
  --base-url https://api.stdcmpt.com/v1 \
  --model standardcompute \
  --api-key-stdin \
  --set-default

jcode --provider-profile standardcompute auth-test --no-tool-smoke

This writes [providers.standardcompute] to ~/.jcode/config.toml and stores the key in jcode's private config directory (for example ~/.config/jcode/provider-standardcompute.env). Add --model-catalog to list every model from /v1/models, then pin one such as anthropic/claude-sonnet-5.5 when a task needs it.

jcode questions

How do I add a custom OpenAI-compatible API to jcode without putting the key in my shell history?

Pipe it in: printf '%s' "$STANDARDCOMPUTE_API_KEY" | jcode provider add standardcompute --base-url https://api.stdcmpt.com/v1 --model standardcompute --api-key-stdin --set-default. jcode writes the profile to ~/.jcode/config.toml and the key to a private env file in its config directory. --api-key-env names an environment variable instead, and --no-api-key is for localhost servers.

Why does a jcode swarm use so many tokens?

The coordinating root of a swarm runs at max effort by default: swarm_root_effort and swarm_deep_root_effort under [agents] in ~/.jcode/config.toml both default to max. Lowering them (for example swarm_root_effort = "low") cuts reasoning on the root without changing the workers' effort. On Standard Compute the model standardcompute also sends each worker step to the cheapest model that can do it, and every session draws on one budget.

Can jcode use an Anthropic-compatible endpoint?

jcode's documented custom profiles (jcode provider add) are OpenAI-compatible, so use https://api.stdcmpt.com/v1. Claude models stay available through it: the router uses them when a step needs them, or pin an exact ID such as anthropic/claude-sonnet-5.5. For Anthropic's own API or a Claude subscription, jcode has its built-in Claude sign-in.