> For the complete documentation index, see [llms.txt](https://docs.darcyiq.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.darcyiq.com/build/integration-overview/aws/bedrock-mantle-integration.md). # AWS Bedrock Mantle Run Darcy's inference on your own AWS account and pick the models behind each tier Bedrock Mantle is Amazon's OpenAI/Anthropic-compatible inference engine. Connecting it points DarcyIQ at a Bedrock project in **your** AWS account, so every model call Darcy makes on your behalf is billed to you — and you choose which models do the work. {% hint style="success" %} This is the recommended way to bring your own models to DarcyIQ. Unlike the older [AWS Bedrock](/build/integration-overview/aws/aws-bedrock-integration.md) option, Mantle lets you pick a specific model for each of Darcy's three workload tiers. {% endhint %} ## Before You Start | Requirement | Detail | | --------------- | -------------------------------------------------------------------- | | **Permissions** | Organization admin or owner | | **AWS region** | `us-east-1`, `us-east-2`, or `us-west-2` | | **AWS access** | Rights to deploy a CloudFormation stack and create a Bedrock API key | ## Setup {% stepper %} {% step %} **Deploy the CloudFormation template** In Darcy, go to **Settings → Organization → Model Routing**, choose **AWS Bedrock Mantle**, and click **Launch CloudFormation template**. The stack creates a Bedrock project in your account, tagged with DarcyIQ's AWS Partner ID so usage is attributed correctly. Deploy it into one of the three supported regions above. {% endstep %} {% step %} **Copy the Project ID** When the stack finishes, copy the **ProjectId** output. It looks like `proj_xxxxxxxx`. {% endstep %} {% step %} **Create a Bedrock API key** In the AWS Console, go to **Amazon Bedrock → API keys** and create a key scoped to the project you just created. {% endstep %} {% step %} **Connect it in Darcy** Back in **Model Routing → AWS Bedrock Mantle**, fill in the three fields and click **Connect**. | Field | Value | | ---------------------- | --------------------------------------- | | **Bedrock API Key** | The bearer token you just created | | **AWS Region** | The region hosting your project | | **Bedrock Project ID** | The `proj_` value from the stack output | | {% endstep %} | | {% step %} **Choose your models** Under **Model selection**, pick a model for each tier and click **Save models**. Your choices apply to the whole organization. {% endstep %} {% endstepper %} ## Choosing Models for Each Tier Darcy routes work to one of three tiers depending on what she's doing. You decide which model backs each one. | Tier | Used for | Default | | -------------- | -------------------------------------------------- | -------------------- | | **Smart** | General chat, agents, and most everyday reasoning | OpenAI GPT-5.6 Terra | | **Fast** | Quick tasks like titles, summaries, and enrichment | OpenAI GPT-5.6 Luna | | **Analytical** | Heavy analysis such as scoping and deep reasoning | OpenAI GPT-5.6 Sol | ### Available Models | Model | Suited to | Regions | | ------------------------------ | ----------------- | ------------------------ | | **OpenAI GPT-5.6 Sol** | Analytical, Smart | `us-east-1`, `us-east-2` | | **OpenAI GPT-5.6 Terra** | Smart, Analytical | All supported regions | | **OpenAI GPT-5.6 Luna** | Fast, Smart | All supported regions | | **OpenAI GPT-5.5** | Smart, Analytical | All supported regions | | **Anthropic Claude Opus 5** | Analytical, Smart | All supported regions | | **Anthropic Claude Sonnet 5** | Smart, Analytical | All supported regions | | **Anthropic Claude Haiku 4.5** | Fast, Smart | All supported regions | {% hint style="info" %} The dropdown for each tier only lists models that suit that workload and are available in your region. GPT-5.6 Sol, for example, won't appear if your project is in `us-west-2`. {% endhint %} Darcy handles the plumbing behind each choice — OpenAI models are called through Mantle's OpenAI-compatible endpoint and Claude models through its Anthropic-compatible one. You pick a model by name; there's nothing else to configure. ## What Changes Once You Connect * **Inference is billed to your AWS account**, through your Bedrock project. * **Darcy behaves the same.** Chat, agents, and every other feature work as before. * **Your model choices apply organization-wide**, not per user. {% hint style="warning" %} **There is no fallback to DarcyIQ's models while Mantle is connected.** If your credentials expire, your key is revoked, or a model errors, the request fails rather than quietly running on our infrastructure. Bringing your own models means Darcy honors that boundary. If you need to fall back, disconnect the integration. {% endhint %} ## One Routing Option at a Time Mantle and the older [AWS Bedrock](/build/integration-overview/aws/aws-bedrock-integration.md) integration can't both be active. If Mantle is connected, the AWS Bedrock card shows as **Unavailable** with a note to disconnect Mantle first. Disconnecting either one returns your organization to DarcyIQ's managed models. ## Troubleshooting | Problem | Cause and fix | | --------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `Bedrock Mantle is not available in ...` | Your project is in a region Darcy doesn't support. The CloudFormation template offers more regions than the integration accepts — redeploy into `us-east-1`, `us-east-2`, or `us-west-2` | | **Model Routing isn't in my settings** | It's limited to organization admins and owners | | **A model disappeared from the dropdown** | It isn't offered in your project's region, or it isn't suited to that tier. Darcy falls back to the tier default | | **Requests started failing after connecting** | Check that your API key is valid and scoped to the project you entered. There's no automatic fallback while Mantle is active | | **Can't find it in Integrations** | Model routing lives under **Settings → Organization → Model Routing**, not the integrations directory | ## Related * [AWS Bedrock](/build/integration-overview/aws/aws-bedrock-integration.md) — the earlier IAM-credential integration * [Organization Management](/settings-and-configuration/organization-management.md) — the rest of your organization's settings --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.darcyiq.com/build/integration-overview/aws/bedrock-mantle-integration.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.