### Quick-start Evaluation Install with Helm

Source: https://deltallm.readthedocs.io/en/latest/deployment/kubernetes

Install DeltaLLM for evaluation using a bundled PostgreSQL and Redis. This is the simplest setup for a first-time install.

```bash
helm install deltallm deltallm/deltallm \
  --version <chart-version> \
  --namespace deltallm \
  --create-namespace \
  -f https://deltawi.github.io/deltallm/values-eval-<chart-version>.yaml \
  --set secret.values.masterKey="$DELTALLM_MASTER_KEY" \
  --set secret.values.saltKey="$DELTALLM_SALT_KEY" \
  --set-string env[0].name=PLATFORM_BOOTSTRAP_ADMIN_EMAIL \
  --set-string env[0].value=admin@example.com \
  --set-string env[1].name=PLATFORM_BOOTSTRAP_ADMIN_PASSWORD \
  --set-string env[1].value='ChangeMe123!'

```

--------------------------------

### Copy Example Configuration File

Source: https://deltallm.readthedocs.io/en/latest/configuration

Use this command to copy the example configuration file to your project for local modification.

```bash
cp config.example.yaml config.yaml

```

--------------------------------

### Install Backend Dependencies with pip

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Create a virtual environment and install backend dependencies using pip and the requirements.txt file.

```bash
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

```

--------------------------------

### Copy Example Config and Set Environment Variable

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Copy the example configuration file and set the DELTALLM_CONFIG_PATH environment variable to point to it.

```bash
cp config.example.yaml config.yaml
export DELTALLM_CONFIG_PATH=./config.yaml

```

--------------------------------

### Start the DeltaLLM Backend API

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Run this command to start the backend API server. It will be accessible at http://localhost:8000.

```bash
uv run uvicorn src.main:app --host 0.0.0.0 --port 8000 --reload
```

--------------------------------

### Install Backend Dependencies with uv

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Use uv to install development dependencies, leveraging the provided uv.lock file for reproducible builds.

```bash
uv sync --dev

```

--------------------------------

### Start Single Instance Deployment

Source: https://deltallm.readthedocs.io/en/latest/getting-started/docker

Starts a single instance of DeltaLLM using Docker Compose. Ensure config.yaml is edited with your settings.

```bash
# Edit config.yaml with your API keys and settings

docker compose --profile single up -d --build

```

--------------------------------

### Start the DeltaLLM Admin UI

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Navigate to the 'ui' directory and run these commands to start the admin UI development server. It will be accessible at http://localhost:5000.

```bash
cd ui
npm ci
npm run dev
```

--------------------------------

### Example of Forwarded Header

Source: https://deltallm.readthedocs.io/en/latest/admin-ui/mcp

An example demonstrating how an authorization header would be prefixed for forwarding through DeltaLLM's MCP.

```text
x-deltallm-mcp-github-authorization: Bearer ...
```

--------------------------------

### Example Success Payload for Tools List

Source: https://deltallm.readthedocs.io/en/latest/api/mcp

This is an example of a successful response when listing available tools, showing the structure of tool information returned.

```JSON
{
  "jsonrpc": "2.0",
  "id": 2,
  "result": {
    "tools": [
      {
        "name": "docs.search",
        "description": "Search docs",
        "inputSchema": {
          "type": "object",
          "properties": {
            "query": {"type": "string"}
          },
          "required": ["query"]
        }
      }
    ]
  }
}
```

--------------------------------

### Start Docker Compose Deployment

Source: https://deltallm.readthedocs.io/en/latest/deployment/docker

Command to start the DeltaLLM application in detached mode using Docker Compose.

```bash
docker compose up -d
```

--------------------------------

### Start DeltaLLM Gateway

Source: https://deltallm.readthedocs.io/en/latest/getting-started/quickstart

Starts the DeltaLLM backend server. Redis can optionally be started for distributed caching and rate limiting.

```bash
# Optional: start Redis for distributed caching and rate limiting
redis-server --daemonize yes

python -m uvicorn src.main:app --host 0.0.0.0 --port 8000 --reload

```

--------------------------------

### MCP Tool Namespace Example

Source: https://deltallm.readthedocs.io/en/latest/features/mcp

DeltaLLM namespaces MCP tools using the format 'server_key.tool_name'. This example shows how a 'search' tool on the 'docs' server is exposed.

```text
server_key.tool_name
```

```text
docs.search
```

--------------------------------

### Chat Example with MCP Tool

Source: https://deltallm.readthedocs.io/en/latest/api/mcp

Use this example to send a chat completion request that includes an MCP tool definition. Ensure the 'tools' parameter is correctly formatted with 'type', 'server', 'allowed_tools', and 'require_approval'.

```curl
curl http://localhost:8000/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [
      {
        "role": "user",
        "content": "Use the docs MCP search tool to look up DeltaLLM."
      }
    ],
    "tools": [
      {
        "type": "mcp",
        "server": "docs",
        "allowed_tools": ["search"],
        "require_approval": "never"
      }
    ],
    "tool_choice": "required"
  }'


```

--------------------------------

### Start High Availability Deployment

Source: https://deltallm.readthedocs.io/en/latest/getting-started/docker

Starts a high-availability deployment with two DeltaLLM instances behind an Nginx load balancer using Docker Compose.

```bash
docker compose --profile ha up -d --build

```

--------------------------------

### Example of Forwarded Authorization Header

Source: https://deltallm.readthedocs.io/en/latest/features/mcp

This specific example illustrates how to forward an authorization header to an MCP server, using 'github' as the server key and 'authorization' as the header name.

```text
x-deltallm-mcp-github-authorization: Bearer ...

```

--------------------------------

### Production Install with External Services

Source: https://deltallm.readthedocs.io/en/latest/deployment/kubernetes

Install DeltaLLM in production mode using an external PostgreSQL and Redis. This requires editing the downloaded `values-production.yaml` file to point to existing secrets.

```bash
helm install deltallm deltallm/deltallm \
  --version <chart-version> \
  --namespace deltallm \
  --create-namespace \
  -f values-production.yaml \
  --set secret.existingSecret=deltallm-app-secrets

```

--------------------------------

### Example Rate Limit Configuration

Source: https://deltallm.readthedocs.io/en/latest/features/rate-limiting

This JSON object shows an example configuration for various rate limits, including requests per minute (rpm), requests per hour (rph), requests per day (rpd), tokens per minute (tpm), and tokens per day (tpd). Use this to set burst tolerance and cost control limits.

```json
{
  "rpm_limit": 60,
  "rph_limit": 500,
  "rpd_limit": 2000,
  "tpm_limit": 100000,
  "tpd_limit": 1000000
}
```

--------------------------------

### Responses Example with MCP Tool

Source: https://deltallm.readthedocs.io/en/latest/api/mcp

This example demonstrates how to make a responses request that incorporates an MCP tool. The 'tools' field should specify the tool's 'type', 'server', 'allowed_tools', and 'require_approval'.

```curl
curl http://localhost:8000/v1/responses \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "input": "Search the docs for DeltaLLM.",
    "tools": [
      {
        "type": "mcp",
        "server": "docs",
        "allowed_tools": ["search"],
        "require_approval": "never"
      }
    ],
    "tool_choice": "required"
  }'


```

--------------------------------

### Example DeltaLLM Deployment Configuration

Source: https://deltallm.readthedocs.io/en/latest/configuration/models

This configuration demonstrates setting up a vLLM gateway deployment with custom upstream authentication headers. Ensure `auth_header_format` contains the `{api_key}` placeholder.

```yaml
model_list:
  - model_name: support-vllm
    deltallm_params:
      provider: vllm
      model: vllm/meta-llama/Llama-3.1-8B-Instruct
      api_key: os.environ/VLLM_GATEWAY_KEY
      api_base: https://vllm.example/v1
      auth_header_name: X-API-Key
      auth_header_format: "{api_key}"
    model_info:
      mode: chat

```

--------------------------------

### Prepare Input File for Batch Requests

Source: https://deltallm.readthedocs.io/en/latest/features/batching

Create a JSONL file where each line represents a batch request. Examples show formats for embeddings and chat completions.

```json
{"custom_id": "doc-1", "url": "/v1/embeddings", "body": {"model": "text-embedding-3-small", "input": "DeltaLLM is an LLM gateway"}}
{"custom_id": "doc-2", "url": "/v1/embeddings", "body": {"model": "text-embedding-3-small", "input": "It supports async batch processing"}}
{"custom_id": "doc-3", "url": "/v1/embeddings", "body": {"model": "text-embedding-3-small", "input": "Batching reduces cost for high-volume workloads"}}

```

```json
{"custom_id": "chat-1", "url": "/v1/chat/completions", "body": {"model": "gpt-4o-mini", "messages": [{"role": "user", "content": "Summarize DeltaLLM in one sentence."}]}}
{"custom_id": "chat-2", "url": "/v1/chat/completions", "body": {"model": "gpt-4o-mini", "messages": [{"role": "user", "content": "Write a short support reply."}]}}

```

--------------------------------

### Install Presidio-enabled DeltaLLM Image

Source: https://deltallm.readthedocs.io/en/latest/deployment/kubernetes

Install the Presidio-enabled variant of DeltaLLM for evaluation. This uses the same eval values file but specifies a different image tag.

```bash
helm install deltallm deltallm/deltallm \
  --version <chart-version> \
  --namespace deltallm \
  --create-namespace \
  -f https://deltawi.github.io/deltallm/values-eval-<chart-version>.yaml \
  --set secret.values.masterKey="$DELTALLM_MASTER_KEY" \
  --set secret.values.saltKey="$DELTALLM_SALT_KEY" \
  --set-string env[0].name=PLATFORM_BOOTSTRAP_ADMIN_EMAIL \
  --set-string env[0].value=admin@example.com \
  --set-string env[1].name=PLATFORM_BOOTSTRAP_ADMIN_PASSWORD \
  --set-string env[1].value='ChangeMe123!' \
  --set image.tag=v<chart-version>-presidio

```

--------------------------------

### First Model Deployment Example

Source: https://deltallm.readthedocs.io/en/latest/configuration/models

This snippet defines a public model name `gpt-4o-mini` and configures it to use an OpenAI deployment with specified API key, base URL, and timeout.

```yaml
model_list:
  - model_name: gpt-4o-mini
    deltallm_params:
      provider: openai
      model: openai/gpt-4o-mini
      api_key: os.environ/OPENAI_API_KEY
      api_base: https://api.openai.com/v1
      timeout: 60

```

--------------------------------

### Example Inline Deployment Payload

Source: https://deltallm.readthedocs.io/en/latest/admin-ui/models

This JSON payload demonstrates how to configure an inline deployment for a vLLM provider, including custom authentication headers.

```json
{
  "model_name": "support-vllm",
  "deltallm_params": {
    "provider": "vllm",
    "model": "vllm/meta-llama/Llama-3.1-8B-Instruct",
    "api_key": "gateway-key",
    "api_base": "https://vllm.example/v1",
    "auth_header_name": "X-API-Key",
    "auth_header_format": "{api_key}"
  },
  "model_info": {
    "mode": "chat"
  }
}
```

--------------------------------

### Install Presidio Engine from Source

Source: https://deltallm.readthedocs.io/en/latest/features/guardrails

Installs the Presidio engine for local development from source. This command adds the necessary dependencies for the full Presidio engine.

```bash
uv sync --dev --extra guardrails-presidio

```

--------------------------------

### Install DeltaLLM with Helm

Source: https://deltallm.readthedocs.io/en/latest/deployment/kubernetes

Install or upgrade the DeltaLLM Helm chart using the evaluation values file and custom generated keys. This command deploys DeltaLLM with bundled PostgreSQL and Redis.

```bash
helm upgrade --install deltallm ./helm \
  --namespace deltallm \
  --create-namespace \
  -f helm/values-eval.yaml \
  --set secret.values.masterKey="$DELTALLM_MASTER_KEY" \
  --set secret.values.saltKey="$DELTALLM_SALT_KEY"

```

--------------------------------

### GET /health/readiness - Readiness Probe

Source: https://deltallm.readthedocs.io/en/latest/api/health

This endpoint checks if all dependencies are connected and the gateway is ready to serve traffic. It returns `200 OK` when ready, and `503 Service Unavailable` otherwise.

```http
GET /health/readiness
```

--------------------------------

### Bootstrap Model Deployments from Config

Source: https://deltallm.readthedocs.io/en/latest/configuration/models

Enables seeding the database with model definitions from `config.yaml`. This is typically used for initial setup. After the first seed, this should be set back to `false`.

```yaml
general_settings:
  model_deployment_source: db_only
  model_deployment_bootstrap_from_config: true

```

--------------------------------

### Create Virtual API Key

Source: https://deltallm.readthedocs.io/en/latest/features/authentication

Create a virtual API key for applications. This key can be scoped and limited, for example, by setting a maximum budget. Requires authentication with the master key.

```bash
curl -X POST http://localhost:8000/ui/api/keys \
  -H "Authorization: Bearer YOUR_MASTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "key_name": "production-app",
    "team_id": "team-production",
    "max_budget": 50.00
  }'

```

--------------------------------

### ElevenLabs Text-to-Speech Deployment (Inline Credentials)

Source: https://deltallm.readthedocs.io/en/latest/configuration/models

Configure an ElevenLabs text-to-speech deployment using inline credentials. This setup routes calls through DeltaLLM's OpenAI-compatible endpoints to ElevenLabs' native API.

```yaml
model_list:
  - model_name: elevenlabs-tts
    deltallm_params:
      provider: elevenlabs
      model: elevenlabs/eleven_multilingual_v2
      api_key: os.environ/ELEVENLABS_API_KEY
      api_base: https://api.elevenlabs.io/v1
    model_info:
      mode: audio_speech
      default_params:
        voice_id: 21m00Tcm4TlvDq8ikWAM
        output_format: mp3_44100_128

```

--------------------------------

### ElevenLabs Speech-to-Text Deployment (Inline Credentials)

Source: https://deltallm.readthedocs.io/en/latest/configuration/models

Configure an ElevenLabs speech-to-text deployment using inline credentials. This setup routes calls through DeltaLLM's OpenAI-compatible endpoints to ElevenLabs' native API.

```yaml
model_list:
  - model_name: elevenlabs-stt
    deltallm_params:
      provider: elevenlabs
      model: elevenlabs/scribe_v2
      api_key: os.environ/ELEVENLABS_API_KEY
      api_base: https://api.elevenlabs.io/v1
    model_info:
      mode: audio_transcription
      default_params:
        timestamps_granularity: word
        diarize: true

```

--------------------------------

### Minimal Starter Configuration

Source: https://deltallm.readthedocs.io/en/latest/configuration/general

This example shows a minimal starter configuration for general settings, sourcing sensitive information like master keys, salt keys, database URLs, and admin credentials from environment variables. It also sets session timeouts and model deployment sources.

```yaml
general_settings:
  master_key: os.environ/DELTALLM_MASTER_KEY
  salt_key: os.environ/DELTALLM_SALT_KEY
  database_url: os.environ/DATABASE_URL
  redis_url: os.environ/REDIS_URL
  platform_bootstrap_admin_email: os.environ/PLATFORM_BOOTSTRAP_ADMIN_EMAIL
  platform_bootstrap_admin_password: os.environ/PLATFORM_BOOTSTRAP_ADMIN_PASSWORD
  auth_session_ttl_hours: 12
  model_deployment_source: db_only
  model_deployment_bootstrap_from_config: true
  governance_notifications_enabled: false
  budget_notifications_enabled: false
  key_lifecycle_notifications_enabled: false

```

--------------------------------

### Initialize Prisma and Apply Database Schema

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Use these commands to generate Prisma client artifacts, fetch binaries, and apply the schema to your local database.

```bash
uv run prisma generate --schema=./prisma/schema.prisma
uv run prisma py fetch
uv run prisma db push --schema=./prisma/schema.prisma
```

--------------------------------

### Build and Serve UI from Backend

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Build the frontend UI and serve it alongside the backend API for a single-process local preview. This is useful for local development and testing.

```bash
cd ui
npm run build
cd ..
uv run uvicorn src.main:app --host 0.0.0.0 --port 5000

```

--------------------------------

### General Settings for Quick Path

Source: https://deltallm.readthedocs.io/en/latest/configuration

Configure essential connection and deployment settings for a simplified runtime. Secrets should be managed via environment variables.

```yaml
general_settings:
  master_key: os.environ/DELTALLM_MASTER_KEY
  salt_key: os.environ/DELTALLM_SALT_KEY
  database_url: os.environ/DATABASE_URL
  redis_url: os.environ/REDIS_URL
  model_deployment_source: db_only
  model_deployment_bootstrap_from_config: true

```

--------------------------------

### Recommended Admin UI Login Variables

Source: https://deltallm.readthedocs.io/en/latest/getting-started/docker

Configure these variables to set up an initial platform admin account for logging into the Admin UI.

```yaml
PLATFORM_BOOTSTRAP_ADMIN_EMAIL=admin@example.com
PLATFORM_BOOTSTRAP_ADMIN_PASSWORD=ChangeMe123!
```

--------------------------------

### Model List Response Example

Source: https://deltallm.readthedocs.io/en/latest/api/proxy

This is an example of the JSON response when listing available models. It shows the structure including the object type, and a list of model details.

```json
{
  "object": "list",
  "data": [
    {
      "id": "gpt-4o-mini",
      "object": "model",
      "owned_by": "deltallm"
    }
  ]
}
```

--------------------------------

### Readiness Probe

Source: https://deltallm.readthedocs.io/en/latest/api/health

Checks that all dependencies are connected. Returns `200 OK` when ready to serve traffic, `503 Service Unavailable` otherwise.

```APIDOC
## GET /health/readiness

### Description
Checks that all dependencies are connected. Returns `200 OK` when ready to serve traffic, `503 Service Unavailable` otherwise.

### Method
GET

### Endpoint
/health/readiness
```

--------------------------------

### Required Environment Variables

Source: https://deltallm.readthedocs.io/en/latest/getting-started/docker

These are the essential environment variables required for the starter config.yaml. Ensure DELTALLM_MASTER_KEY and DELTALLM_SALT_KEY are unique and not placeholders.

```yaml
DELTALLM_MASTER_KEY=sk-your-master-key
DELTALLM_SALT_KEY=your-random-salt-key
OPENAI_API_KEY=sk-your-openai-key
```

--------------------------------

### GET /auth/login

Source: https://deltallm.readthedocs.io/en/latest/api/admin

Initiates the Single Sign-On (SSO) login flow.

```APIDOC
## GET /auth/login

### Description
Initiates the Single Sign-On (SSO) login flow.

### Method
GET

### Endpoint
/auth/login
```

--------------------------------

### POST /auth/mfa/enroll/start

Source: https://deltallm.readthedocs.io/en/latest/api/admin

Initiates the Multi-Factor Authentication (MFA) enrollment process.

```APIDOC
## POST /auth/mfa/enroll/start

### Description
Starts the Multi-Factor Authentication (MFA) enrollment process.

### Method
POST

### Endpoint
/auth/mfa/enroll/start
```

--------------------------------

### GET /auth/sso-config

Source: https://deltallm.readthedocs.io/en/latest/api/admin

Retrieves the Single Sign-On (SSO) configuration settings.

```APIDOC
## GET /auth/sso-config

### Description
Retrieves the Single Sign-On (SSO) configuration settings.

### Method
GET

### Endpoint
/auth/sso-config
```

--------------------------------

### GET /auth/me

Source: https://deltallm.readthedocs.io/en/latest/api/admin

Inspects the current session state for the authenticated user.

```APIDOC
## GET /auth/me

### Description
Inspects the current session state for the authenticated user.

### Method
GET

### Endpoint
/auth/me
```

--------------------------------

### Quick Path Router Settings

Source: https://deltallm.readthedocs.io/en/latest/configuration/router

A minimal configuration for most first deployments, setting basic routing strategy, retries, timeouts, and cooldown.

```yaml
router_settings:
  routing_strategy: simple-shuffle
  num_retries: 1
  retry_after: 1
  timeout: 600
  cooldown_time: 60
  allowed_fails: 0

```

--------------------------------

### Get Deployment

Source: https://deltallm.readthedocs.io/en/latest/api/admin

Retrieves details for a specific model deployment by its ID.

```APIDOC
## GET /ui/api/models/{deployment_id}

### Description
Get one deployment

### Method
GET

### Endpoint
/ui/api/models/{deployment_id}

### Parameters
#### Path Parameters
- **deployment_id** (string) - Required - The ID of the deployment to retrieve.
```

--------------------------------

### GET /auth/callback

Source: https://deltallm.readthedocs.io/en/latest/api/admin

Completes the Single Sign-On (SSO) login process after redirection.

```APIDOC
## GET /auth/callback

### Description
Completes the Single Sign-On (SSO) login process after redirection.

### Method
GET

### Endpoint
/auth/callback
```

--------------------------------

### Client Request with Tags

Source: https://deltallm.readthedocs.io/en/latest/features/routing

Example of a client request specifying tags for tag-based routing.

```json
{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "hello"}],
  "metadata": {"tags": ["eu"]}
}

```

--------------------------------

### Get Batch Status and Progress

Source: https://deltallm.readthedocs.io/en/latest/features/batching

Fetches the current status and progress details of a specific batch job.

```APIDOC
## GET /v1/batches/{batch_id}

### Description
Get batch status and progress.

### Method
GET

### Endpoint
/v1/batches/{batch_id}

### Parameters
#### Path Parameters
- **batch_id** (string) - Required - The ID of the batch to retrieve status for.
```

--------------------------------

### Get One Event

Source: https://deltallm.readthedocs.io/en/latest/features/audit-log

Retrieves a single audit event by its ID, including any stored payload records.

```APIDOC
## GET /ui/api/audit/events/{event_id}

### Description
Retrieves a specific audit event by its unique identifier, including associated payload records if available.

### Method
GET

### Endpoint
/ui/api/audit/events/{event_id}

### Parameters
#### Path Parameters
- **event_id** (string) - Required - The unique identifier of the audit event to retrieve.

### Response
#### Success Response (200)
- **event_id** (string) - Unique identifier for the event.
- **occurred_at** (string) - Timestamp when the event occurred.
- **organization_id** (string) - The ID of the organization.
- **actor_type** (string) - The type of the actor (e.g., USER, SYSTEM).
- **actor_id** (string) - The ID of the actor.
- **action** (string) - The type of action performed.
- **resource_type** (string) - The type of resource affected.
- **resource_id** (string) - The ID of the resource affected.
- **request_id** (string) - The ID of the request.
- **correlation_id** (string) - The correlation ID for the event.
- **status** (string) - The status of the action (e.g., SUCCESS, FAILURE).
- **latency_ms** (integer) - The latency of the operation in milliseconds.
- **input_tokens** (integer) - Number of input tokens.
- **output_tokens** (integer) - Number of output tokens.
- **error_type** (string) - Type of error, if any.
- **error_code** (string) - Error code, if any.
- **metadata** (object) - Additional metadata associated with the event.
- **content_stored** (boolean) - Indicates if content was stored for this event.
- **payloads** (object) - Stored payload records associated with the event, if any.
```

--------------------------------

### Create Deployment Referencing a Named Credential

Source: https://deltallm.readthedocs.io/en/latest/admin-ui/named-credentials

This example shows how to create a deployment that uses a previously created named credential. When using a named credential, provider-specific connection fields should be omitted from the deployment payload.

```bash
curl http://localhost:4002/ui/api/models \
  -H "Authorization: Bearer $DELTALLM_MASTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model_name": "support-vllm",
    "named_credential_id": "cred-123",
    "deltallm_params": {
      "provider": "vllm",
      "model": "vllm/meta-llama/Llama-3.1-8B-Instruct"
    },
    "model_info": {
      "mode": "chat"
    }
  }'

```

--------------------------------

### Generate Security Keys

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Generate unique DELTALLM_MASTER_KEY and DELTALLM_SALT_KEY using Python and OpenSSL. These are required for DeltaLLM to start.

```bash
export DELTALLM_MASTER_KEY="$(python3 -c 'import secrets; print(\"sk-\" + secrets.token_hex(20) + \"A1\")')"
export DELTALLM_SALT_KEY="$(openssl rand -hex 32)"

```

--------------------------------

### Apply Prisma Schema

Source: https://deltallm.readthedocs.io/en/latest/getting-started/docker

Applies the Prisma schema to the database on startup for the DeltaLLM container.

```bash
prisma db push --schema=./prisma/schema.prisma --accept-data-loss

```

--------------------------------

### Get Named Credential

Source: https://deltallm.readthedocs.io/en/latest/admin-ui/named-credentials

Retrieves the details of a specific named credential, including its usage count and linked deployments.

```APIDOC
## GET /ui/api/named-credentials/{credential_id}

### Description
Retrieves the details of a specific named credential.

### Method
GET

### Endpoint
/ui/api/named-credentials/{credential_id}

### Parameters
#### Path Parameters
- **credential_id** (string) - Required - The ID of the credential to retrieve.

### Response
#### Success Response (200)
- **credential_id** (string) - The unique identifier for the credential.
- **name** (string) - The name of the credential.
- **provider** (string) - The provider of the service.
- **connection_config** (object) - Configuration details for the connection (sensitive fields redacted).
- **credentials_present** (boolean) - Indicates if credentials are set.
- **usage_count** (integer) - The number of times the credential has been used.
- **linked_deployments** (array) - A list of deployments linked to this credential.

#### Response Example
```json
{
  "credential_id": "cred-123",
  "name": "vLLM Shared Gateway",
  "provider": "vllm",
  "connection_config": {
    "api_key": "***REDACTED***",
    "api_base": "https://vllm.example/v1",
    "auth_header_name": "X-API-Key",
    "auth_header_format": "{api_key}"
  },
  "credentials_present": true,
  "usage_count": 5,
  "linked_deployments": ["deploy-abc", "deploy-xyz"]
}
```
```

--------------------------------

### Example Response Shape for Named Credential Creation

Source: https://deltallm.readthedocs.io/en/latest/admin-ui/named-credentials

This is the expected JSON response structure when a named credential is successfully created.

```json
{
  "credential_id": "cred-123",
  "name": "vLLM Shared Gateway",
  "provider": "vllm",
  "connection_config": {
    "api_key": "***REDACTED***",
    "api_base": "https://vllm.example/v1",
    "auth_header_name": "X-API-Key",
    "auth_header_format": "{api_key}"
  },
  "credentials_present": true,
  "usage_count": 0
}

```

--------------------------------

### Export DeltaLLM Base URL and Master Key

Source: https://deltallm.readthedocs.io/en/latest/getting-started/mcp-quickstart

Set environment variables for the DeltaLLM base URL and your master key. Use `http://localhost:4002` for Docker quick starts or `http://localhost:8000` for local backends.

```bash
export BASE="http://localhost:4002"
export MASTER_KEY="YOUR_MASTER_KEY"

```

--------------------------------

### Build and Run Deltallm Docker Image

Source: https://deltallm.readthedocs.io/en/latest/getting-started/docker

Builds the Deltallm Docker image and runs a container with specified ports, environment variables, and volume mounts. Ensure the database is reachable and keys are set.

```bash
docker build -t deltallm .
docker run -p 4002:4000 \
  -e DATABASE_URL="postgresql://..." \
  -e DELTALLM_MASTER_KEY="sk-your-key" \
  -e DELTALLM_SALT_KEY="your-salt-key" \
  -v ./config.yaml:/app/config/config.yaml:ro \
  deltallm

```

--------------------------------

### Prometheus Scrape Configuration

Source: https://deltallm.readthedocs.io/en/latest/features/observability

Example Prometheus configuration to scrape metrics from DeltaLLM. Ensure Prometheus is configured to access the `/metrics` endpoint.

```yaml
scrape_configs:
  - job_name: deltallm
    scrape_interval: 15s
    static_configs:
      - targets: ["localhost:8000"]
    metrics_path: /metrics

```

--------------------------------

### Configure Environment Variables for Local Development

Source: https://deltallm.readthedocs.io/en/latest/getting-started/installation

Set essential environment variables for database connection, security, API keys, and admin credentials. Ensure DELTALLM_MASTER_KEY and DELTALLM_SALT_KEY are unique and not placeholder values.

```bash
export DATABASE_URL="postgresql://postgres:postgres@localhost:5432/deltallm"
export DELTALLM_MASTER_KEY="sk-local-1234567890abcdefghijklmnop"
export DELTALLM_SALT_KEY="$(openssl rand -hex 32)"
export OPENAI_API_KEY="sk-your-provider-key"
export PLATFORM_BOOTSTRAP_ADMIN_EMAIL="admin@example.com"
export PLATFORM_BOOTSTRAP_ADMIN_PASSWORD="ChangeMe123!"

```