API Documentation
General
- POSTAdd LLM Provider
- POSTAdd Ollama Models
- POSTChange Node Name
- GETGet Default Embedding Model
- POSTSet Default Embedding Model
- GETDownload File from Inbox
- GETAPI Health Check
- POSTInitial Registration
- GETCheck Pristine State
- GETList Files in Inbox
- GETGet Local Processing Preference
- POSTSet Local Processing Preference
- POSTModify LLM Provider
- GETGet Public Keys
- POSTRemove LLM Provider
- GETScan Ollama Models
- POSTStop LLM Process
- GETGet Supported Embedding Models
- POSTUpdate Supported Embedding Models
Jobs
- POSTAdd File to Inbox
- GETGet All Inboxes
- GETGet Available LLM Providers
- POSTChange Job LLM Provider
- POSTCreate Files Inbox
- POSTCreate Job
- GETGet Job Configuration
- GETGet Job Scope
- POSTSend Job Message
- POSTGet Last Messages
- POSTGet Last Messages with Branches
- POSTRetry Message
- POSTUpdate Job Configuration
- POSTUpdate Job Scope
- POSTUpdate Smart Inbox Name
Subscriptions
- POSTGet Available Shared Items
- POSTGet Open Shared Items
- POSTCreate Shareable Folder
- POSTGenerate Free Subscription Links
- POSTGet Latest Notifications
- POSTList My Subscribers
- POSTList My Subscriptions
- POSTGet Notifications Before Timestamp
- POSTSubscribe to Shared Folder
- POSTUnshare Folder
- POSTUnsubscribe
- POSTUpdate Shareable Folder
Tool Offerings
Vector File System
Get Job Scope
curl --request GET \
--url http://127.0.0.1:9550/v2/get_job_scope{
"local_vrkai": [
{
"vrkai": {
"metadata": {},
"resource": {
"Document": {
"created_datetime": "2023-12-25",
"data_tag_index": {
"index": {}
},
"description": "<string>",
"distribution_info": {
"datetime": "2023-12-25",
"origin": {
"Uri": "<string>"
}
},
"embedding_model_used_string": "<string>",
"embeddings": [
{
"id": "<string>",
"vector": [
123
]
}
],
"keywords": {
"keyword_list": [
"<string>"
],
"keywords_embedding": {
"embedding": {
"id": "<string>",
"vector": [
123
]
},
"model_used": "<string>"
}
},
"last_written_datetime": "2023-12-25",
"merkle_root": "<string>",
"metadata_index": {
"index": {}
},
"name": "<string>",
"node_count": 1,
"nodes": [
{
"content": {
"Text": "<string>"
},
"data_tag_names": [
"<string>"
],
"id": "<string>",
"last_written_datetime": "2023-12-25",
"merkle_hash": "<string>",
"metadata": {}
}
],
"resource_base_type": "Document",
"resource_embedding": {
"id": "<string>",
"vector": [
123
]
},
"resource_id": "<string>",
"source": {
"Standard": {
"FileRef": {
"file_name": "<string>",
"file_type": {
"Document": "Pdf"
},
"text_chunking_strategy": "V1"
}
}
}
}
},
"sfm": {
"map": {},
"source_files_count": 1
},
"version": "V1"
}
}
],
"local_vrpack": [
{
"vrpack": {
"embedding_models_used": {},
"folder_count": 1,
"metadata": {},
"name": "<string>",
"resource": {
"Document": {
"created_datetime": "2023-12-25",
"data_tag_index": {
"index": {}
},
"description": "<string>",
"distribution_info": {
"datetime": "2023-12-25",
"origin": {
"Uri": "<string>"
}
},
"embedding_model_used_string": "<string>",
"embeddings": [
{
"id": "<string>",
"vector": [
123
]
}
],
"keywords": {
"keyword_list": [
"<string>"
],
"keywords_embedding": {
"embedding": {
"id": "<string>",
"vector": [
123
]
},
"model_used": "<string>"
}
},
"last_written_datetime": "2023-12-25",
"merkle_root": "<string>",
"metadata_index": {
"index": {}
},
"name": "<string>",
"node_count": 1,
"nodes": [
{
"content": {
"Text": "<string>"
},
"data_tag_names": [
"<string>"
],
"id": "<string>",
"last_written_datetime": "2023-12-25",
"merkle_hash": "<string>",
"metadata": {}
}
],
"resource_base_type": "Document",
"resource_embedding": {
"id": "<string>",
"vector": [
123
]
},
"resource_id": "<string>",
"source": {
"Standard": {
"FileRef": {
"file_name": "<string>",
"file_type": {
"Document": "Pdf"
},
"text_chunking_strategy": "V1"
}
}
}
}
},
"version": "V1",
"vrkai_count": 1
}
}
],
"network_folders": [
{
"external_identity": {
"full_name": "<string>",
"node_name": "<string>",
"profile_name": "<string>",
"subidentity_name": "<string>",
"subidentity_type": "Agent"
},
"name": "<string>",
"path": {
"path_ids": [
"<string>"
]
}
}
],
"vector_fs_folders": [
{
"name": "<string>",
"path": {
"path_ids": [
"<string>"
]
}
}
],
"vector_fs_items": [
{
"name": "<string>",
"path": {
"path_ids": [
"<string>"
]
},
"source": {
"Standard": {
"FileRef": {
"file_name": "<string>",
"file_type": {
"Document": "Pdf"
},
"text_chunking_strategy": "V1"
}
}
}
}
]
}Fetch the defined scope or boundaries of a particular job.
Query Parameters
Job ID to retrieve scope for
Response
Job's scope which includes both Local entries (vrkai stored locally only in job) and VecFS entries (source/vector resource stored in the FS, accessible to all jobs)
A Scope Entry for a local VRKai that only lives in the Job's scope (not in the VectorFS & thus not available to other jobs)
Represents a parsed VRKai file with a BaseVectorResource, and optional SourceFileMap.
To save as a file or transfer the VRKai, call one of the prepare_as_ methods. To parse from a file/transfer, use the from_ methods.
The list of base/core VectorResource types which are fully composable within one another
A VectorResource which uses an internal numbered/ordered list data model, thus providing an ideal interface for document-like content such as PDFs, epubs, web content, written works, and more.
Info about where the source data was acquired from, and when it was originally distributed
A struct that represents a Vector Embedding with an added string id.
A struct which holds a Vector Resource's keywords/optional keywords embedding
x >= 0Represents a Vector Resource Node which holds a unique id, one of the types of NodeContent, metadata, and other internal relevant data.
Enum used for VectorResources to self-attest their base type.
CustomUnsupported(s) allows for devs to implement custom VectorResources that fulfill the trait,
but which aren't composable with any of the base resources (we are open to PRs for adding new base types as well).
Document A struct that represents a Vector Embedding with an added string id.
Information about the source content a Vector Resource came from
V1 A map which stores SourceFiles based on VRPaths within a VectorResource. A SourceFile at root represents the single source file for the whole VR.
A Scope Entry for a local VRPack that only lives in the Job's scope (not in the VectorFS & thus not available to other jobs)
Represents a parsed VRPack file, which contains a Map Vector Resource that holds a tree structure of folders & encoded VRKai nodes.
In other words, a .vrpack file is akin to a "compressed archive" of internally held VRKais with folder structure preserved.
Of note, VRPacks are not compressed at the top level because the VRKais held inside already are. This improves performance for large VRPacks.
To save as a file or transfer the VRPack, call one of the encode_as_ methods. To parse from a file/transfer, use the from_ methods.
x >= 0x >= 0VRPack metadata enables users to add extra info that may be needed for unique use cases
The list of base/core VectorResource types which are fully composable within one another
A VectorResource which uses an internal numbered/ordered list data model, thus providing an ideal interface for document-like content such as PDFs, epubs, web content, written works, and more.
Info about where the source data was acquired from, and when it was originally distributed
A struct that represents a Vector Embedding with an added string id.
A struct which holds a Vector Resource's keywords/optional keywords embedding
x >= 0Represents a Vector Resource Node which holds a unique id, one of the types of NodeContent, metadata, and other internal relevant data.
Enum used for VectorResources to self-attest their base type.
CustomUnsupported(s) allows for devs to implement custom VectorResources that fulfill the trait,
but which aren't composable with any of the base resources (we are open to PRs for adding new base types as well).
Document A struct that represents a Vector Embedding with an added string id.
Information about the source content a Vector Resource came from
V1 x >= 0A Scope Entry for a FSFolder that (potentially) exists on another node's VectorFS (if your node has perms). Unsupported currently, struct added for future compatibility.
Agent, Device A Scope Entry for a FSFolder saved in the VectorFS.
A Scope Entry for a FSItem saved in the VectorFS.
A path inside of a Vector Resource to a Node which exists somewhere in the hierarchy. Internally the path is made up of an ordered list of Node ids (Int-holding strings for Docs, any string for Maps).
Information about the source content a Vector Resource came from
Type that acts as a reference to external file/content/data
What text chunking strategy was used to create this VR from the source file. This is required for performing content validation/that it matches the VR nodes. TODO: Think about how to make this more explicit/specific and future support
V1 Was this page helpful?
curl --request GET \
--url http://127.0.0.1:9550/v2/get_job_scope{
"local_vrkai": [
{
"vrkai": {
"metadata": {},
"resource": {
"Document": {
"created_datetime": "2023-12-25",
"data_tag_index": {
"index": {}
},
"description": "<string>",
"distribution_info": {
"datetime": "2023-12-25",
"origin": {
"Uri": "<string>"
}
},
"embedding_model_used_string": "<string>",
"embeddings": [
{
"id": "<string>",
"vector": [
123
]
}
],
"keywords": {
"keyword_list": [
"<string>"
],
"keywords_embedding": {
"embedding": {
"id": "<string>",
"vector": [
123
]
},
"model_used": "<string>"
}
},
"last_written_datetime": "2023-12-25",
"merkle_root": "<string>",
"metadata_index": {
"index": {}
},
"name": "<string>",
"node_count": 1,
"nodes": [
{
"content": {
"Text": "<string>"
},
"data_tag_names": [
"<string>"
],
"id": "<string>",
"last_written_datetime": "2023-12-25",
"merkle_hash": "<string>",
"metadata": {}
}
],
"resource_base_type": "Document",
"resource_embedding": {
"id": "<string>",
"vector": [
123
]
},
"resource_id": "<string>",
"source": {
"Standard": {
"FileRef": {
"file_name": "<string>",
"file_type": {
"Document": "Pdf"
},
"text_chunking_strategy": "V1"
}
}
}
}
},
"sfm": {
"map": {},
"source_files_count": 1
},
"version": "V1"
}
}
],
"local_vrpack": [
{
"vrpack": {
"embedding_models_used": {},
"folder_count": 1,
"metadata": {},
"name": "<string>",
"resource": {
"Document": {
"created_datetime": "2023-12-25",
"data_tag_index": {
"index": {}
},
"description": "<string>",
"distribution_info": {
"datetime": "2023-12-25",
"origin": {
"Uri": "<string>"
}
},
"embedding_model_used_string": "<string>",
"embeddings": [
{
"id": "<string>",
"vector": [
123
]
}
],
"keywords": {
"keyword_list": [
"<string>"
],
"keywords_embedding": {
"embedding": {
"id": "<string>",
"vector": [
123
]
},
"model_used": "<string>"
}
},
"last_written_datetime": "2023-12-25",
"merkle_root": "<string>",
"metadata_index": {
"index": {}
},
"name": "<string>",
"node_count": 1,
"nodes": [
{
"content": {
"Text": "<string>"
},
"data_tag_names": [
"<string>"
],
"id": "<string>",
"last_written_datetime": "2023-12-25",
"merkle_hash": "<string>",
"metadata": {}
}
],
"resource_base_type": "Document",
"resource_embedding": {
"id": "<string>",
"vector": [
123
]
},
"resource_id": "<string>",
"source": {
"Standard": {
"FileRef": {
"file_name": "<string>",
"file_type": {
"Document": "Pdf"
},
"text_chunking_strategy": "V1"
}
}
}
}
},
"version": "V1",
"vrkai_count": 1
}
}
],
"network_folders": [
{
"external_identity": {
"full_name": "<string>",
"node_name": "<string>",
"profile_name": "<string>",
"subidentity_name": "<string>",
"subidentity_type": "Agent"
},
"name": "<string>",
"path": {
"path_ids": [
"<string>"
]
}
}
],
"vector_fs_folders": [
{
"name": "<string>",
"path": {
"path_ids": [
"<string>"
]
}
}
],
"vector_fs_items": [
{
"name": "<string>",
"path": {
"path_ids": [
"<string>"
]
},
"source": {
"Standard": {
"FileRef": {
"file_name": "<string>",
"file_type": {
"Document": "Pdf"
},
"text_chunking_strategy": "V1"
}
}
}
}
]
}