Python¶

The local engine¶

from mimir import Mimir

model = Mimir.from_pretrained("Mythologic/MIMIR-1")

Argument	Default	Does
`model`	`Mythologic/MIMIR-1`	a Hugging Face Hub id or a path to a local release directory
`revision`	the revision this package version pins	a specific Hub commit or tag
`device`	`auto`	`cpu`, `cuda`, or CUDA when a compatible GPU is available
`variant`	the variant listed for the device	`fp32` (CPU) or `fp16` (CUDA)
`policy`	the release policy	path to a custom policy JSON from `mimir calibrate`
`cache_dir`	the Hub cache	where the model files are stored
`offline`	`False`	load only from the local cache, with no network access
`allow_unsigned`	`False`	load a local release directory that has no Sigstore signature

mimir.Mimir requires mimir-decisions[local] (CPU) or mimir-decisions[local-gpu] (CUDA). Both extras install the same onnxruntime module, so keep only one in any given environment. mimir doctor reports the active runtime and names any conflict.

Before loading the ONNX session, from_pretrained checks the pinned revision, verifies the manifest's Sigstore signature against the release identity of abderahmane-ai/mimir, verifies each file's SHA-256 against the manifest, and checks the ONNX graph against its operator allowlist and signature. Nothing is read until every check passes.

For offline use, fetch once with mimir download and then load with offline=True (or set HF_HUB_OFFLINE=1).

model.info() reports the model id, revision, variant, runtime fingerprint, certified risk levels, and input limits. model.count_tokens(context, spec) sizes a request before sending it.

The engine is safe to share across threads.

The HTTP client¶

from mimir import MimirClient

remote = MimirClient("https://mimir.internal", api_key="...")
remote.choose("...", "Which team?", options=["billing", "security"])

MimirClient implements the same interface as Mimir — the same methods, the same signatures — so code, decision tools, and framework adapters accept either without modification. It requires only the base install.

api_key defaults to the MIMIR_API_KEY environment variable. Connection errors, timeouts, and 429, 502, 503, 504, and 529 responses are retried with exponential backoff that honours Retry-After. Any other error response raises a ServerResponseError subclass that keeps the HTTP status and the response body.

Use MimirClient as a context manager to ensure the underlying connection pool is released:

with MimirClient("https://mimir.internal") as remote:
    result = remote.choose("...", "Which team?", options=["billing", "security"])

Async and batches¶

Every method has an async counterpart: adecide, adecide_many, achoose, ayes_no, averify, arank, arate, aestimate.

decide_many and adecide_many take a list of (context, spec) pairs and batch them by token length. On the local engine, requests that share the same risk level and alpha are grouped into a single engine call.

Decision tools¶

from mimir import Choice

route_ticket = model.tool(
    "route_ticket",
    Choice("Which team should handle this ticket?", options=["billing", "security"]),
    description="Route a support ticket to the team that owns it.",
)
route_ticket("My card was charged twice")
route_ticket.input_schema, route_ticket.output_schema

A decision tool binds a spec to a name so the caller supplies only the context. The HTTP server, the MCP server, and every framework adapter expose decision tools. input_schema and output_schema are JSON Schema objects describing the tool's expected input and output.

Types without the engine¶

from mimir import Choice, Context, ChoiceResult loads only the data models, which depend on Pydantic alone. A project that constructs requests or reads results without running the model can depend on the base install and never touch ONNX Runtime.