[heydryft]

Adds Runcrate as a built-in OpenAI-compatible provider, mirroring the STACKIT entry everywhere it appears. Runcrate serves open models (DeepSeek V4.1 Flash, GLM-5.3, Qwen3.8 Flash, gpt-oss 120B) at https://api.runcrate.ai/v1.

• jcode-provider-metadata/src/catalog.rs: RUNCRATE_PROFILE (RUNCRATE_API_KEY, runcrate.env, default deepseek-ai/DeepSeek-V4.1-Flash) and RUNCRATE_LOGIN_PROVIDER, added to both arrays (sizes bumped)
• jcode-base: key redaction in message.rs, env mapping in auth/external.rs, pricing id in model_pricing.rs, fallback model list in provider_catalog.rs
• src/cli/provider_init.rs / args.rs: --provider runcrate
• openrouter_tests.rs: profile + default model case

Testing:

• cargo test -p jcode-provider-metadata -p jcode-provider-openrouter-runtime: pass
• cargo test -p jcode --lib provider: 54 passed
• cargo test -p jcode-base --lib: same failures as master on this machine (auth::transfer and a few env-dependent ones), none new
• Live: RUNCRATE_API_KEY=... jcode --provider runcrate run --json "Reply with exactly: RUNCRATE_OK" returns "provider": "runcrate", "model": "deepseek-ai/DeepSeek-V4.1-Flash", "text": "RUNCRATE_OK"

Disclosure: I work on Runcrate.

[greptile-apps]

Confidence Score: 5/5

[Medium impact] No blocking merge-safety concerns were identified in the reviewed Runcrate integration.

Summary

Adds Runcrate as a built-in OpenAI-compatible provider.

• Jcode can now use Runcrate as a built-in provider.

No actionable issues were found. The shared menu order retains both providers, and the DeepSeek default remains eligible for reasoning replay. No blocking merge-safety concerns were identified within the reviewed scope. Live Runcrate availability and vendor-side model support were not tested.

_{Reviews (1) · Last reviewed commit: "feat(providers): add Runcrate as an Open..." · Reviewed by Greptile}