vrr/openclaw
2
0
Fork 0
mirror of https://github.com/openclaw/openclaw.git synced 2026-10-03 09:39:25 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 30
openclaw/extensions/nvidia
History
Peter Steinberger c6e6380c5e
refactor(providers): deslop provider glue (#162897) 
Reuse existing requirements, WebSocket byte conversion, normalization, auth failure, diagnostic truncation, stream type, and catalog grouping owners across provider adapters and agent glue. Remove private forwarding layers while preserving provider policy and lifecycle boundaries.

Reject nonfinite Vercel catalog prices and per-million conversion overflow using the existing fallback. Regression cases cover numeric and scaled overflow.

Production reduction: 427 lines. Full provider and core suites, changed checks, plugin contracts, import boundaries, SDK API comparison, and zero-cycle checks validate the refactor; fresh isolated review found no actionable issue.
2026-10-01 18:03:40 +00:00
..
assets improve(plugins): give bundled logos consistent white icon tiles (#155259) 2026-09-23 19:09:26 -07:00
api.ts
index.test.ts test(cron,providers): remove low-value tests (batch d045) (#159165) 2026-09-26 21:38:25 +00:00
index.ts refactor(providers): deslop provider plugins fourth pass (#161471) 2026-09-29 19:11:06 -07:00
nvidia.live.test.ts fix(models): preserve DeepInfra and NVIDIA catalogs on discovery failure (#139749) 2026-09-06 11:16:56 +05:30
onboard.test.ts test(cron,providers): remove low-value tests (batch d045) (#159165) 2026-09-26 21:38:25 +00:00
onboard.ts refactor(providers): deslop provider plugins fourth pass (#161471) 2026-09-29 19:11:06 -07:00
openclaw.plugin.json feat(plugins): define package-owned categories (#142710) 2026-09-09 19:05:40 -07:00
package.json chore(release): close out 2026.9.7 on main (#161587) 2026-09-29 22:39:14 -07:00
provider-catalog.test.ts fix(models): keep shared catalog requests alive for remaining callers (#157191) 2026-09-27 19:51:12 +08:00
provider-catalog.ts refactor(providers): deslop provider glue (#162897) 2026-10-01 18:03:40 +00:00
README.md feat: show declared plugin capabilities and setup guides (#157956) 2026-09-25 16:43:53 -07:00
tsconfig.json

README.md

NVIDIA

Use NVIDIA-hosted models through NVIDIA's OpenAI-compatible API. OpenClaw provides model setup and a catalog of supported chat models, including available NVIDIA and third-party models.

Get started

Create an NVIDIA API key and run:

openclaw onboard --auth-choice nvidia-api-key

You can also supply NVIDIA_API_KEY in the Gateway's environment. Browse models with openclaw models list --provider nvidia and select an available model for your agent.

This plugin calls NVIDIA's hosted service; it does not install local GPU models. Catalog entries and supported inputs depend on NVIDIA's current inventory.

See the NVIDIA guide for authentication and model selection.