Together AI Provider Setup Guide

Together AI offers a unified API for reaching popular open-source models such as Llama, DeepSeek, Kimi, and others. OpenClaw ships it as the together provider.

Property	Value
Provider	`together` (alias: `togetherai`)
Auth	`TOGETHER_API_KEY`
API	OpenAI-compatible
Base URL	`https://api.together.xyz/v1`

Getting started

Get an API key

Generate an API key from api.together.ai/settings/api-keys.

Run onboarding

openclaw onboard --auth-choice together-api-key

Set a default model

{
  agents: {
    defaults: {
      model: {
        primary: "together/zai-org/GLM-5.2",
      },
    },
  },
}

Non-interactive example

openclaw onboard --non-interactive --accept-risk --skip-health \
  --mode local \
  --auth-choice together-api-key \
  --together-api-key "$TOGETHER_API_KEY"

Note

During onboarding, together/zai-org/GLM-5.2 is configured as the default chat model.

Built-in catalog

Prices are listed in USD per million tokens.

Model ref	Name	Input	Context	Max output	Cost (in/out)	Notes
`together/meta-llama/Llama-3.3-70B-Instruct-Turbo`	Llama 3.3 70B Instruct Turbo	text	131,072	8,192	1.04 / 1.04	General model
`together/zai-org/GLM-5.2`	GLM 5.2 FP4	text	262,144	131,072	1.40 / 4.40	Default model

Video generation

Video generation is also registered by the bundled together plugin, which works through the shared video_generate tool.

Property	Value
Default video model	`Wan-AI/wan2.7-t2v`
Other models	`Wan-AI/wan2.7-i2v`, `minimax/hailuo-02`
Modes	text-to-video; image-to-video only with `Wan-AI/wan2.7-i2v` (single first-frame image)
Duration	Wan 2.7: 2-10 seconds in OpenClaw; Hailuo duration support is model-specific
Supported parameters	`size` (see below); `aspectRatio`/`resolution` are not read

When Wan 2.7 is used, OpenClaw maps size onto Together's resolution and ratio fields. The sizes that are accepted are documented in Together's Wan 2.7 guide:

Aspect ratio	720P size	1080P size
16:9	`1280x720`	`1920x1080`
9:16	`720x1280`	`1080x1920`
1:1	`960x960`	`1440x1440`
4:3	`1104x832`	`1648x1248`
3:4	`832x1104`	`1248x1648`

Any other Wan size is rejected before the request is submitted. Leaving out size lets Together apply its defaults. For Hailuo, pixel width and height are still passed through from size. Wan image input travels as media.frame_images together with frame: "first", rather than as a style reference in media.reference_images.

To make Together your default video provider:

{
  agents: {
    defaults: {
      mediaModels: {
        video: {
          primary: "together/Wan-AI/wan2.7-t2v",
        },
      },
    },
  },
}

Tip

Shared tool parameters, how providers are selected, and failover behavior are covered in Video generation.

Prompt caching

Matching prompt prefixes are reused automatically by Together. To improve routing between turns, OpenClaw includes a stable prompt_cache_key on native Together requests. Depending on the model's response format, cache usage is read from usage.prompt_tokens_details.cached_tokens and usage.cached_tokens. Reuse works best when the model, tool definitions, reasoning settings, and earlier messages stay unchanged. See Together prompt caching.

Environment note

When the Gateway is run as a daemon (launchd/systemd), confirm that TOGETHER_API_KEY can be seen by that process (set it in ~/.openclaw/.env, for instance, or through env.shellEnv).

Warning

A key exported only in your interactive shell will not be visible to gateway processes managed by a daemon. For persistent availability, use ~/.openclaw/.env or the env.shellEnv config.

Troubleshooting

Check that your key works: openclaw models list --provider together
When models fail to show up, make sure the API key is set in the environment your Gateway process actually reads.
Model refs take the form together/<model-id>.

Model providers, Rules for providers, model refs, and how failover works.
Video generation, Parameters shared by the video generation tool, plus how a provider is chosen.
Configuration reference, The complete config schema, provider settings included.
Together AI, The Together AI dashboard, API documentation, and pricing.