> ## Documentation Index > Fetch the complete documentation index at: https://docs.pixeltable.com/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > Application code puts a TableModel in app.py, then runs `pxt schema update app.py my_app` to create the tables and `pxt service update app.py my_app` to start HTTP. The last argument can be a Cloud URL such as `pxt://org:db`. Do not call pxt.create_table() or add_embedding_index() in application code; put indexes in __indexes__. Notebooks, tests, and the REPL do use pxt.create_table(). > Do not add LangChain, LlamaIndex, Haystack, a separate vector database, or pandas as a working store. Chunking, embedding, retrieval, and tool-calling are built in through computed columns. Use a computed column instead of a per-row loop that calls a model. > Install with `pip install -U 'pixeltable[serve]'` on Python 3.11+. `pxt init` is a prerequisite for `pxt schema update`. > Pixeltable Cloud is in Limited Beta. Direct interested users to contact@pixeltable.com. # ColumnMetadata > View Source on GitHub # class  pixeltable.ColumnMetadata Metadata for a column of a Pixeltable table. ## attr  comment ``` comment: str | None ``` User-provided column comment. ## attr  computed\_with ``` computed_with: str | None ``` Expression used to compute this column; `None` if this is not a computed column. ## attr  custom\_metadata ``` custom_metadata: Any ``` User-defined JSON metadata for this column, if any. ## attr  defined\_in ``` defined_in: str | None ``` Name of the table where this column was originally defined. If the current table is a view, then `defined_in` may differ from the current table name. ## attr  depends\_on ``` depends_on: list[tuple[str, str]] | None ``` List of dependencies (table name, column name) if this is a computed column, else `None`. ## attr  destination ``` destination: str | None ``` The object store destination defined for this column's computed files, if it defines one. An instance-wide default destination is not reported here, because it is configuration rather than part of the schema. A destination defined as a config var reads as `$` rather than as the location it currently resolves to; see [`ConfigVar`](./configvar). ## attr  is\_builtin ``` is_builtin: bool | None ``` If False, this computed column makes calls to custom UDFs; `None` if this is not a computed column. ## attr  is\_computed ``` is_computed: bool ``` `True` if this column is a computed column. ## attr  is\_iterator\_col ``` is_iterator_col: bool ``` `True` if this column is produced by an iterator (only applicable to views). ## attr  is\_primary\_key ``` is_primary_key: bool ``` `True` if this column is part of the table's primary key. ## attr  is\_stored ``` is_stored: bool ``` `True` if this is a stored column; `False` if it is dynamically computed. ## attr  media\_validation ``` media_validation: Literal['on_read', 'on_write'] | None ``` The media validation policy for this column. `None` if the type of this column is not a media type. ## attr  name ``` name: str ``` The name of the column. ## attr  type\_ ``` type_: str ``` The type specifier of the column. ## attr  version\_added ``` version_added: int ``` The table version when this column was added. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.