Requests

`LoraAdapterGeneration`

Select one exact immutable LoRA adapter generation by its 64-character ID.

Field	Type
`generation`	`str`

`ChatCompletionRequest`

A ChatCompletionRequest represents a request sent to the mistral.rs engine. It encodes information about input data, sampling, and how to return the response.

The messages type is as follows: (for normal chat completion, for chat completion with images, pretemplated prompt)

Agent permission fields:

agent_permission: AgentPermission.Auto, .Ask, or .Deny. Applies to server-executed agent actions such as code execution, shell, web search, file tools, callbacks, and external tool dispatch.
agent_approval_callback: called when agent_permission=AgentPermission.Ask with an AgentToolApproval. Return True, False, or AgentToolApprovalDecision.

See agent permissions for the shared CLI, HTTP, Python, and Rust behavior.

reasoning_effort accepts off, low, medium, high, and xhigh; none aliases off. Values are trimmed and case-insensitive. If both reasoning controls are omitted, thinking is enabled with no selected effort. Contradictory enable_thinking and reasoning_effort values raise ValueError.

Field	Type	Default
`messages`	`list[dict[str, str]] \| list[dict[str, list[dict[str, str \| dict[str, str]]]]] \| str`	required
`model`	`str`	required
`logprobs`	`bool`	`False`
`n_choices`	`int`	`1`
`logit_bias`	`dict[int, float] \| None`	`None`
`top_logprobs`	`int \| None`	`None`
`max_tokens`	`int \| None`	`None`
`presence_penalty`	`float \| None`	`None`
`frequency_penalty`	`float \| None`	`None`
`repetition_penalty`	`float \| None`	`None`
`stop_seqs`	`list[str] \| None`	`None`
`temperature`	`float \| None`	`None`
`top_p`	`float \| None`	`None`
`top_k`	`int \| None`	`None`
`stream`	`bool`	`False`
`grammar`	`str \| None`	`None`
`grammar_type`	`str \| None`	`None`
`min_p`	`float \| None`	`None`
`tool_schemas`	`list[str] \| None`	`None`
`tool_choice`	`ToolChoice \| None`	`None`
`dry_multiplier`	`float \| None`	`None`
`dry_base`	`float \| None`	`None`
`dry_allowed_length`	`int \| None`	`None`
`dry_sequence_breakers`	`list[str] \| None`	`None`
`web_search_options`	`WebSearchOptions \| None`	`None`
`enable_thinking`	`bool \| None`	`None`
`truncate_sequence`	`bool`	`False`
`reasoning_effort`	`Literal['off', 'none', 'low', 'medium', 'high', 'xhigh'] \| None`	`None`
`max_tool_rounds`	`int \| None`	`None`
`tool_dispatch_url`	`str \| None`	`None`
`enable_code_execution`	`bool`	`False`
`enable_shell`	`bool`	`False`
`shell_skills`	`list[ShellSkillMount] \| None`	`None`
`agent_permission`	`AgentPermission \| None`	`None`
`agent_approval_callback`	`Callable[[AgentToolApproval], bool \| AgentToolApprovalDecision] \| None`	`None`
`code_execution_permission`	`CodeExecutionPermission \| None`	`None`
`session_id`	`str \| None`	`None`
`files`	`list[RequestedFile] \| None`	`None`
`input_files`	`list[InputFile] \| None`	`None`
`ignore_eos`	`bool`	`False`
`adapter`	`str \| LoraAdapterGeneration \| None`	`None (keyword-only)`

`CompletionRequest`

A CompletionRequest represents a request sent to the mistral.rs engine. It encodes information about input data, sampling, and how to return the response.

Field	Type	Default
`prompt`	`str`	required
`model`	`str`	required
`best_of`	`int`	`1`
`echo_prompt`	`bool`	`False`
`presence_penalty`	`float \| None`	`None`
`frequency_penalty`	`float \| None`	`None`
`repetition_penalty`	`float \| None`	`None`
`logit_bias`	`dict[int, float] \| None`	`None`
`max_tokens`	`int \| None`	`None`
`n_choices`	`int`	`1`
`stop_seqs`	`list[str] \| None`	`None`
`temperature`	`float \| None`	`None`
`top_p`	`float \| None`	`None`
`suffix`	`str \| None`	`None`
`top_k`	`int \| None`	`None`
`grammar`	`str \| None`	`None`
`grammar_type`	`str \| None`	`None`
`min_p`	`float \| None`	`None`
`tool_schemas`	`list[str] \| None`	`None`
`tool_choice`	`ToolChoice \| None`	`None`
`dry_multiplier`	`float \| None`	`None`
`dry_base`	`float \| None`	`None`
`dry_allowed_length`	`int \| None`	`None`
`dry_sequence_breakers`	`list[str] \| None`	`None`
`truncate_sequence`	`bool`	`False`
`ignore_eos`	`bool`	`False`
`adapter`	`str \| LoraAdapterGeneration \| None`	`None (keyword-only)`

`EmbeddingRequest`

An EmbeddingRequest represents a request to compute embeddings for the provided input text.

Field	Type	Default
`input`	`str \| list[str] \| list[int] \| list[list[int]]`	required
`truncate_sequence`	`bool`	`False`

Generated from mistralrs-pyo3/mistralrs.pyi.