# Answer from your documents

The agent reads your documents and cites the file and page, instead of answering from memory.

The agent answers from documents it opens with its tools. Pick where they live:

| Documents | Put them | The agent finds them with |
| --- | --- | --- |
| A small fixed set (policies, a price list) | The recipe's `files/` folder | `grep` and `read` |
| Different per task, up to 100 files | Uploads attached to the task | `read`, `pdftotext` |
| More than 100, or changing (a wiki, a data room) | Your search or storage MCP server | That server's tools |

**You need:** a recipe you can change ([Build](/build)).

## 1. Ship a small set with the recipe

Put text or Markdown (not PDF: `grep` can't search it) in a subfolder of
`files/`, and give the agent the file tools.

```yaml filename="agents/agent.yaml"
tools: [read, grep, find, ls, bash]
```

```md filename="SYSTEM.md"
Answer only from the documents in /workspace/files/policies/. Search them with grep,
read the matching file, and cite its name. If nothing matches, say so.
```

`files/policies/refunds.md` arrives read-only at `/workspace/files/policies/refunds.md`.

## 2. Attach documents per task

```bash
introspection tasks create --runtime support-agent -e production \
  --file @./contract.pdf --prompt "Does this contract allow early termination?"
```

A task takes up to 100 attachments, each up to 25 MB
([Send files in and out](/patterns/files)). `pdftotext` and `pdftoppm` are in
every runtime.

## 3. Cite the page

Tell the agent to extract one page at a time, so every quote has a page number:

```md filename="SYSTEM.md"
For PDFs, read page by page with `pdftotext -f N -l N file.pdf -`. Cite every
figure and quote as (file name, p. N). Never cite a page you did not read.
```

To require citations rather than hope for them, make `sources` a required
field of a [submit tool](/patterns/structured-output#require-it-dont-just-measure-it).

## 4. Review a large set

Past 100 files, or for a findings grid across a data room:

- **Search:** bind a search or storage server holding the set ([MCP](/build/mcp)).
- **Fan out:** start one task per document, in waves ([batch](/patterns/trigger#4-run-a-batch)), each writing one small JSON output, then merge them in your code.
- **Split in one task:** the main agent hands documents to helpers, up to 64 running at once ([Split work across agents](/patterns/multi-agent)).

## Verify

```bash
introspection local -p "What is our refund window? Cite the file."
introspection conversations get <task-id> -o json \
  --query "items[?attributes.gen_ai.operation.name=='execute_tool'].attributes.gen_ai.tool.name"
```

Run `introspection local` from the recipe directory. Expect an answer naming a
file, and `grep`, `read`, `bash`, or search calls in the cloud task.

Or ask your agent: "Make support-agent answer only from files/policies and cite file and page."

## Limits

| Not supported | Do this instead |
| --- | --- |
| A built-in vector index | Your search server (step 4). |
| More than 100 attachments per task | Search, or fan out (step 4). |
| Scanned PDFs as text | `pdftoppm -png`, then `read` the images with a vision model. |
| One shared search key for every end user | Authorize an [application](/identity/applications#tell-a-partner-who-the-end-user-is) on the endpoint and filter on the JWT `sub`. |

Answer from your documents

The agent reads your documents and cites the file and page, instead of answering from memory.

The agent answers from documents it opens with its tools. Pick where they live:

Documents	Put them	The agent finds them with
A small fixed set (policies, a price list)	The recipe’s `files/` folder	`grep` and `read`
Different per task, up to 100 files	Uploads attached to the task	`read`, `pdftotext`
More than 100, or changing (a wiki, a data room)	Your search or storage MCP server	That server’s tools

You need: a recipe you can change (Build).

1. Ship a small set with the recipe

Put text or Markdown (not PDF: grep can’t search it) in a subfolder of files/, and give the agent the file tools.

agents/agent.yaml


tools: [read, grep, find, ls, bash]

SYSTEM.md


Answer only from the documents in /workspace/files/policies/. Search them with grep,
read the matching file, and cite its name. If nothing matches, say so.

files/policies/refunds.md arrives read-only at /workspace/files/policies/refunds.md.

2. Attach documents per task


introspection tasks create --runtime support-agent -e production \
  --file @./contract.pdf --prompt "Does this contract allow early termination?"

A task takes up to 100 attachments, each up to 25 MB (Send files in and out). pdftotext and pdftoppm are in every runtime.

3. Cite the page

Tell the agent to extract one page at a time, so every quote has a page number:

SYSTEM.md


For PDFs, read page by page with `pdftotext -f N -l N file.pdf -`. Cite every
figure and quote as (file name, p. N). Never cite a page you did not read.

To require citations rather than hope for them, make sources a required field of a submit tool.

4. Review a large set

Past 100 files, or for a findings grid across a data room:

Search: bind a search or storage server holding the set (MCP).
Fan out: start one task per document, in waves (batch), each writing one small JSON output, then merge them in your code.
Split in one task: the main agent hands documents to helpers, up to 64 running at once (Split work across agents).

Verify


introspection local -p "What is our refund window? Cite the file."
introspection conversations get <task-id> -o json \
  --query "items[?attributes.gen_ai.operation.name=='execute_tool'].attributes.gen_ai.tool.name"

Run introspection local from the recipe directory. Expect an answer naming a file, and grep, read, bash, or search calls in the cloud task.

Or ask your agent: “Make support-agent answer only from files/policies and cite file and page.”

Limits

Not supported	Do this instead
A built-in vector index	Your search server (step 4).
More than 100 attachments per task	Search, or fan out (step 4).
Scanned PDFs as text	`pdftoppm -png`, then `read` the images with a vision model.
One shared search key for every end user	Authorize an application on the endpoint and filter on the JWT `sub`.