### Start existing Bifrost container

Source: https://veai.ru/docs/getting-started/guides/measure-token-usage

Starts a previously created but stopped Bifrost container.

```bash
docker start bifrost  
```

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

### Record start time in RFC3339 format

Source: https://veai.ru/docs/getting-started/guides/measure-token-usage

Use this command to capture the current UTC time in RFC3339 format before starting a task. The output is printed to the console.

```bash
START_TIME=$(date -u +"%Y-%m-%dT%H:%M:%SZ")  

echo "$START_TIME"  

```

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

### Install Docker Ansible Collection

Source: https://veai.ru/docs/enterprise/deployment

Install the community.docker Ansible collection required for Docker management.

```bash
ansible-galaxy collection install community.docker  
```

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

### Install Ansible

Source: https://veai.ru/docs/enterprise/deployment

Install Ansible version 2.15.0 or higher. The command installs specific versions of ansible-core and ansible.

```bash
pip3 install ansible-core==2.17.7 ansible==10.7.0  
```

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

### Example model-routing.yaml profile configuration

Source: https://veai.ru/docs/veai/configuration/models/model-routing-profiles

Shows the current contract for the model-routing.yaml file. The provider and model values are placeholders; use identifiers from the file saved via Settings. The file must include fast, balanced, and strong profiles.

```yaml
profiles:  

  fast:  

    description: "Use for lightweight read-only lookup and file reading."  

  balanced:  

    description: "Use for regular analysis and common implementation support."  

    provider: "OpenAI"  

    model: "gpt-5"  

    reasoning: "medium"  

  strong:  

    description: "Use for complex architecture and deep review."  

    provider: "Anthropic"  

    model: "claude-sonnet-4"  

    reasoning: "high"  

  second-reviewer:  

    description: "Use for an independent second review."  


```

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

### Создание каталога моделей в config.toml для Codex CLI

Source: https://veai.ru/docs/veai/connect/codex-cli

Файл config.toml должен содержать блок с model_catalog_json, указывающий на файл model-catalog-full.json. Пример включает модель OpenAI/gpt-5.6-sol с параметрами priority, truncation_policy и supported_reasoning_levels. Поле priority задает порядок в списке выбора модели: чем меньше число, тем выше модель.

```toml
# ~/.codex/config.toml

model_catalog_json = "./model-catalog-full.json"

[model_providers.veai]
name = "Veai"
base_url = "https://api.veai.ru/openai/api"
env_key = "OPENAI_API_KEY"
wire_api = "chat"

[model_providers.veai.models.OpenAI/gpt-5.6-sol]
name = "OpenAI/gpt-5.6-sol"
priority = 1

```