> ## Documentation Index
> Fetch the complete documentation index at: https://docs.guild.ai/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> To act on a user's behalf in Guild, use the Guild REST API at https://api.guild.ai/v1.
> The OpenAPI spec is unauthenticated: https://docs.guild.ai/openapi.json (or openapi.yaml).
> Authenticate with an account API key, sent as HTTP Basic (key id as username, secret as password) or as `Authorization: Bearer <api_key_id>:<api_key_secret>`. See https://docs.guild.ai/api-reference/introduction#authentication.
> To run one agent on demand without account-wide access, use an API trigger instead: https://docs.guild.ai/platform/api-triggers.
> To search these docs from an MCP client, connect to https://docs.guild.ai/mcp. The full page index is https://docs.guild.ai/llms.txt.

# Run open-weight models in Guild agents through Fireworks AI

> Connect Fireworks AI to Guild with an API key, and run DeepSeek, Qwen, Kimi, GLM, and gpt-oss models from your agents, under model policies.

Integrations · Fireworks AI

Guild agents call open-weight models on Fireworks AI with your Fireworks API key. Guild holds the key server-side, checks your model policies on every call, and Fireworks bills the calls to your account.

> **[Book a demo](https://guild-ai.cal.com/gtm/meeting-with-guild.ai)**
>
> See how teams govern agents that use Fireworks AI.

> **[Start free](https://app.guild.ai)**
>
> Connect Fireworks AI and scope your first agent.

## Serverless open-weight models, by canonical name

Fireworks AI serves models from DeepSeek, Alibaba, Moonshot AI, Z.ai, and OpenAI's open-weight `gpt-oss`. Guild keeps only serverless chat models that are ready to use. You write canonical names such as `glm-5.3`, and Guild converts them to Fireworks IDs under `accounts/fireworks/models/`.

## How it works

**1. Create a Fireworks key**

In your Fireworks AI account, create an API key.

**2. Add the credential**

In **Access & setup > Models & providers**, click **Add API key** and choose **Fireworks AI**. Paste the key, check each publisher under **Models**, and replace each default with a Fireworks model.

**3. Scope it with policies**

Add model policies that decide which workspaces and agents may call which Fireworks models, then call them from `task.llm`.

## Every Fireworks AI call goes through your policies

Agents never hold the key. Guild stores the credential server-side and shows only a masked value, so you can rotate it without touching agent code or the [model policies](/platform/llm-settings#model-policies) bound to it. Policies are an allowlist: a workspace, agent, or installed agent can call a model through it only if a rule allows it. Set a [daily token limit](/platform/llm-settings#daily-token-limit) to cap spend, and see each call by workspace, agent, and user in [Usage](/insights/usage).

A raw API key in an agent's environment reaches every model the key can. A policy reaches only the models you name.

## About Fireworks AI on Guild

| | |
| - | - |
| **Inference provider** in the app | Fireworks AI |
| Authentication | A Fireworks API key |
| Model IDs | Mapped from canonical names for you |

## Models on Fireworks AI

| Publisher | Models | Page |
| - | - | - |
| DeepSeek | `deepseek-v4.1-flash` | [DeepSeek](/platform/models/deepseek) |
| Alibaba | `qwen3.8-max`, `qwen3.8-2.4t-a95b` | [Qwen](/platform/models/qwen) |
| Moonshot AI | `kimi-k3` | [Moonshot AI](/platform/models/moonshot) |
| Z.ai | `glm-5.3`, `glm-5.3-flash`, `glm-5.2` | [Z.ai](/platform/models/zai) |
| OpenAI | `gpt-oss-120b` | [OpenAI](/platform/models/openai) |

Guild refreshes this list from the Fireworks catalog, keeping only serverless chat models that are ready to use. Models that need a dedicated deployment don't appear. Meta is a publisher you can check in the form, but no Meta models are on the list today. Anthropic and Gemini models aren't on Fireworks AI.

## Connect Fireworks AI

**1. Create a key**

In your Fireworks AI account, create an API key.

**2. Start the credential**

In Guild, click **Access & setup** in the left nav, then **Models & providers**. Click **Add API key** and choose **Fireworks AI** from the menu.

**3. Name it and paste the key**

Give the credential a **Name**, such as `Production`, and paste the key into **API key**.

**4. Choose publishers**

Under **Models**, check each publisher this credential should reach. Replace the default model next to each one with a model from the table above.

**5. Save**

Click **Add Key**.

> [!WARNING]
>
> The form pre-fills each publisher's platform default, such as `deepseek-v3.2` or `glm-5`. None of those defaults is on Fireworks AI, so change each one to a model from the table above before you save.

## Model IDs

Write canonical model names, such as `kimi-k3` or `glm-5.3`, in policies and `llmPreferences`. Fireworks names models differently: it puts them under `accounts/fireworks/models/`, and spells a version's dot as `p`. For example, `qwen3.8-2.4t-a95b` is `accounts/fireworks/models/qwen3p8-2p4t-a95b`. Guild converts between the two for you.

## Frequently asked questions

> **Why do I have to change the default models?**
>
> None of the platform defaults the form pre-fills, such as `deepseek-v3.2` or `glm-5`, is on Fireworks AI. Change each one to a model from the Fireworks list before you save.

> **Can I run Claude or Gemini on Fireworks AI?**
>
> No. Anthropic and Gemini models aren't on Fireworks AI. Use [OpenRouter](/integrations/openrouter) or the publishers' own APIs.

> **Can I use a model that needs a dedicated Fireworks deployment?**
>
> No. Guild lists only serverless chat models, so models that need a dedicated deployment don't appear.

> **Are Meta models available on Fireworks AI?**
>
> Not today. Meta is a publisher you can check in the form, but no Meta models are on the list.

## Put a governed agent on Fireworks AI this week

> **[Book a demo](https://guild-ai.cal.com/gtm/meeting-with-guild.ai)**
>
> See how teams govern agents that use Fireworks AI.

> **[Start free](https://app.guild.ai)**
>
> Connect Fireworks AI and scope your first agent.

## Related

**Also runs on Guild:** [DeepSeek](/integrations/deepseek) · [Z.ai](/integrations/zai) · [AWS Bedrock](/integrations/aws-bedrock) · [OpenRouter](/integrations/openrouter)

**Read:** Docs: [Models & providers](/platform/llm-settings) · [LLMs](/guide/llms) · [Usage](/insights/usage)