DirectAI MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| DIRECTAI_HOME | No | Overrides the configuration directory. Defaults to %USERPROFILE%\.directai\ | |
| DIRECTAI_TOKEN | No | Yandex Direct API token. If not provided, the server uses the token stored in Windows Credential Manager. | |
| DIRECTAI_TEST_GUARD | No | Set to '1' to disable the guard (testing only). Do not set in production. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_accountsB | Кабинеты из кеша discover, активные кампании, остаток баллов. |
| search_actionsC | Поиск действий каталога. mode: read, write или any. |
| describe_actionC | Описание действия: параметры (JSON Schema), пример, ограничения. |
| run_readC | Выполнить действие чтения. Возвращает Markdown-текст. |
| plan_writeC | Подготовить запись: предпросмотр, предупреждения, plan_id. |
| apply_writeB | Применить план. Только после согласия пользователя. |
| get_operation_logC | Последние операции записи из журнала. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 7 tools
Each tool maps to a distinct stage or resource: account listing, catalog search, action introspection, read execution, write planning, write application, and logging. The plan/apply pair is complementary rather than ambiguous, and read/execute vs. catalog search are clearly separated.
All tool names follow a consistent lowercase snake_case verb-first pattern: list_, search_, describe_, run_, plan_, apply_, get_. The naming style is uniform and predictable across the entire set.
Seven tools is well within the ideal range and each tool earns its place in the workflow. The count matches the server's apparent scope: account discovery, action inspection, read execution, and a two-phase write flow.
The toolset covers the main lifecycle well: discover accounts, search and describe actions, run reads, plan writes, apply them, and review the operation log. Minor gaps exist, such as no explicit cancel-plan operation or fetching a single log entry by ID, but agents can work around these.