timesfm-mcp
by chokukil
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TIMESFM_MODEL_ID | No | Hugging Face model id or local model path. | google/timesfm-2.5-200m-pytorch |
| TIMESFM_BATCH_SIZE | No | TimesFM per_core_batch_size; raise only after memory is stable. | 64 |
| TIMESFM_MAX_CONTEXT | No | Maximum context points used by the compiled model. | 1024 |
| TIMESFM_MAX_HORIZON | No | Maximum forecast horizon. | 256 |
| TIMESFM_TORCH_COMPILE | No | Enable PyTorch compile when loading the model. | false |
| TIMESFM_RETURN_BACKCAST | No | Internal default; XReg requests force this to true. | false |
| TIMESFM_NORMALIZE_INPUTS | No | Normalize each input series before forecasting. | true |
| TIMESFM_INFER_IS_POSITIVE | No | Clamp positive-only series to nonnegative outputs. | true |
| TIMESFM_FIX_QUANTILE_CROSSING | No | Enforce monotonic quantiles. | true |
| TIMESFM_FORCE_FLIP_INVARIANCE | No | Enforce sign symmetry. | true |
| TIMESFM_CONTINUOUS_QUANTILE_HEAD | No | Use continuous quantile head for better bands. | true |
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
Server capabilities have not been inspected yet.
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
No tools | |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |