Scopix

Documentation

AgentSearchResource

Access via client.agent_search - AI-powered search without chat sessions

Direct Agent Access

Use this for standalone AI searches. For conversational search with context, use ChatResource instead.

Tip: More Token-Efficient Than Chat

These standalone search methods consume significantly fewer tokens than running searches through agentic chat. If you only need to find images or documents without multi-turn conversation, use these directly.

images()

Execute AI-powered image search with intelligent reasoning

python
async def images(
    query: str,
    *,
    limit: int = 50,  # 1-500
    folder_id: Optional[str] = None,
    image_ids: Optional[list[str]] = None,
) -> ImageSearchAgentResult
Returns: ImageSearchAgentResult - Full results with metadata, summary, and result refs

Example

python
# Search for images
result = await client.agent_search.images(
    "damaged utility poles in residential areas"
)
print(f"Found {result.count} images")
print(f"Summary: {result.summary}")


for img in result.results[:5]:
    print(f"- {img.filename}: score={img.score:.2f}")


# Search within a folder
result = await client.agent_search.images(
    "safety equipment",
    folder_id="folder_abc123"
)

documents()

Execute AI-powered document search with intelligent reasoning

python
async def documents(
    query: str,
    *,
    limit: int = 50,  # 1-500
    document_types: Optional[list[str]] = None,  # pdf, docx, txt
    document_ids: Optional[list[str]] = None,
) -> DocumentSearchAgentResult
Returns: DocumentSearchAgentResult - Chunks with text, scores, and document references

Example

python
# Search documents
result = await client.agent_search.documents(
    "safety inspection procedures",
    document_types=["pdf"]
)


print(f"Found {result.count} sections in {len(result.document_ids)} documents")
for chunk in result.results:
    print(f"{chunk.document_filename} p{chunk.page_numbers}: {chunk.text[:100]}...")

videos()

Execute AI-powered video search with scene-level matching. Returns matched videos and the specific scenes (with start/end timestamps) inside each that correspond to the query.

python
async def videos(
    query: str,
    *,
    limit: int = 50,  # 1-500
    folder_id: Optional[str] = None,
    video_ids: Optional[list[str]] = None,
    domain: Optional[str] = None,  # general, civil_engineering, maps, mining
) -> VideoSearchAgentResult
Returns: VideoSearchAgentResult — Videos with matched_scenes, summary, and result refs

Example

python
result = await client.agent_search.videos(
    "closeup of a fire truck at night"
)
print(f"Found {result.count} videos")
print(f"Summary: {result.summary}")


for video in result.results:
    print(f"- {video.video_filename}  score={video.score:.2f}")
    for scene in video.matched_scenes or []:
        print(f"    scene {scene.scene_index}  "
              f"{scene.time_range_formatted}  {scene.description}")


# Search within specific videos
result = await client.agent_search.videos(
    "sparks during grinding",
    video_ids=["vid_550e8400", "vid_660f9500"],
)

Related

Standalone Agents Guide

How to use search and operation agents

Search Types

ImageSearchAgentResult, DocumentSearchAgentResult, VideoSearchAgentResult