Integration

Progress Agentic RAG integration

Progress Agentic RAG (formerly Nuclia) indexes your documents in a Knowledge Box and answers questions over them. Ask questions answered by an LLM with citations, run semantic, keyword and graph searches, suggest and summarize; create, update and delete resources and their text, link, file, conversation and key-value fields; manage label sets, synonyms, key-value schemas and saved search configurations; and follow processing, import and export status. Not exposed: Knowledge Box creation and account management, TUS and binary uploads, file downloads, the streaming notifications and the predict proxy, and the models configuration and extract or split strategies (their settings can carry your own LLM provider keys). File and link field values are withheld from resource reads, since they can carry a file password or crawl cookies. Authentication is a Knowledge Box API key (created in the dashboard under Advanced, API Keys, with a reader, writer, manager or owner role; it expires after one year by default) sent as Bearer in the X-NUCLIA-SERVICEACCOUNT header. The connection also needs the Knowledge Box id and its zone, the first label of the Knowledge Box endpoint the dashboard shows (zone.dp.progress.cloud, for example europe-1).

Connect Progress Agentic RAG How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Progress Agentic RAG.Built from the 60 Progress Agentic RAG tools below, on the connection you add once.

What you can do with Progress Agentic RAG

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Resource, 8 operations

create_resourceCreate resourcePOSTCreate a resource with its fields (texts, links, files, conversations, key_values) and metadata. Returns uuid; fields are processed asynchronously (see get_processing_status). Key role: writer.
delete_resourceDelete resourceDELETEPermanently delete a resource and all its fields from the Knowledge Box. Key role: writer.
get_resourceGet resourceGETRead one resource by id. Use show to choose what comes back (basic, origin, extra, relations, values, extracted, errors, security). File and link field values are withheld. Key role: reader.
get_resource_by_slugGet resource by slugGETRead one resource by its slug (your own id for it) instead of its uuid; same show options and output as get_resource. File and link field values are withheld. Key role: reader.
list_resourcesList resourcesGETList the resources of the Knowledge Box page by page (page is 0-based, size defaults to 20); pagination.last says when to stop. Key role: reader.
reindex_resourceReindex resourcePOSTRebuild the search index entries of a resource from its already processed data (set reindex_vectors to true to rebuild its vectors too). Key role: writer.
reprocess_resourceReprocess resourcePOSTSend every field of a resource back to processing (extraction and indexing run again). Key role: writer.
update_resourceUpdate resourcePATCHChange a resource's title, summary, slug, metadata or fields; only what you send is changed. Changed fields are reprocessed. Key role: writer.

Kv schema, 5 operations

create_kv_schemaCreate kv schemaPOSTCreate a key-value schema: an id, a description and its fields (each with a key, a type and a description). Key-value fields of resources are then validated against it. Key role: writer.
delete_kv_schemaDelete kv schemaDELETEDelete a key-value schema. This cannot be undone. Key role: writer.
get_kv_schemaGet kv schemaGETRead one key-value schema with its fields. Key role: reader.
list_kv_schemasList kv schemasGETList the key-value schemas of the Knowledge Box (the shapes that key-value fields of resources must follow). Key role: reader.
update_kv_schemaUpdate kv schemaPUTUpdate a key-value schema's description or fields. Send the full field list (read get_kv_schema first): the vendor does not say whether fields left out are kept. Key role: writer.

Search configuration, 5 operations

create_search_configurationCreate search configurationPOSTSave a named search configuration: kind find or ask, and config, an object of the same parameters find_paragraphs_advanced or ask_knowledge_box take. Needs an owner key.
delete_search_configurationDelete search configurationDELETEDelete a saved search configuration. Requests that load it by name will no longer find it. Needs an owner key.
get_search_configurationGet search configurationGETRead one saved search configuration: its kind (find or ask) and its parameters. Key role: reader.
list_search_configurationsList search configurationsGETList the saved search configurations (named sets of find or ask parameters that requests can load by name). Key role: reader.
update_search_configurationUpdate search configurationPATCHUpdate a saved search configuration; kind and config are required. Read get_search_configuration first and send the full config back with your change. Needs an owner key.

Labelset, 4 operations

delete_labelsetDelete labelsetDELETEDelete a label set and its labels. This cannot be undone. Key role: writer.
get_labelsetGet labelsetGETRead one label set with its labels, kind, color and whether several labels may be applied. Key role: reader.
list_labelsetsList labelsetsGETList the label sets of the Knowledge Box. Key role: reader.
set_labelsetSet labelsetPOSTCreate a label set or replace it whole with what you send (title, color, multiple, kind, labels). Read get_labelset first to keep existing labels. Key role: writer.

Resource field, 3 operations

delete_resource_fieldDelete resource fieldDELETEPermanently delete one field (and its extracted data) from a resource. Key role: writer.
get_resource_fieldGet resource fieldGETRead one field of a resource with its extracted data (text, metadata, vectors, question answers) and processing status or errors. The raw field value is withheld. Key role: reader.
reprocess_resource_fieldReprocess resource fieldPOSTSend one field of any type back to processing; for a password-protected file use reprocess_file_field or pass x-file-password here. Key role: writer.

Synonym, 3 operations

delete_custom_synonymsDelete custom synonymsDELETEDelete all custom synonyms of the Knowledge Box. This cannot be undone. Key role: writer.
get_custom_synonymsGet custom synonymsGETRead the Knowledge Box custom synonyms, a map from a term to its list of synonyms (used by keyword search with with_synonyms). Key role: reader.
set_custom_synonymsSet custom synonymsPUTReplace the Knowledge Box custom synonyms with the map you send (term to list of synonyms). Terms left out are removed; read get_custom_synonyms first to keep them. Key role: writer.

Answer, 2 operations

ask_knowledge_boxAsk knowledge boxPOSTAsk a question; the Knowledge Box retrieves matching paragraphs and its LLM writes an answer with optional citations. Returns answer, retrieval_results and learning_id (for send_feedback). Key role: reader.
ask_resourceAsk resourcePOSTAsk a question answered only from one resource; same options and output as ask_knowledge_box, including learning_id for send_feedback. Key role: reader.

Catalog, 2 operations

list_catalogList catalogGETList resources by title and metadata (no semantic search), with filters, facets, processing status and sort; paged by page_number and page_size (max 200). Key role: reader.
search_catalogSearch catalogPOSTList resources like list_catalog but with the JSON body form, which accepts a filter_expression object and a sort object; paged by page_number and page_size (max 200). Key role: reader.

Conversation message, 2 operations

append_conversation_messagesAppend conversation messagesPUTAppend messages to an existing conversation field of a resource (items = array of message objects, each with a unique ident). Key role: writer.
delete_conversation_messageDelete conversation messageDELETEDelete one message from a conversation field by its ident. Meant for occasional deletions: the vendor says frequent use on one conversation may degrade read performance. Key role: writer.

Entities group, 2 operations

get_entities_groupGet entities groupGETRead one entities group (a named list of entities such as PERSON or ORG), with its entities. Key role: reader.
list_entities_groupsList entities groupsGETList the entities groups of the Knowledge Box; set show_entities to true to include each group's entities. Key role: reader.

File field, 2 operations

reprocess_file_fieldReprocess file fieldPOSTSend one file field of a resource back to processing, optionally with the password of a protected file. Key role: writer.
set_file_fieldSet file fieldPUTAdd or replace a file field on a resource; file carries the bytes as base64 (or a uri). The file is then processed and indexed. Key role: writer.

Paragraph, 2 operations

find_paragraphsFind paragraphsGETSemantic and keyword search returning the best matching paragraphs grouped by resource and field (the recommended search). top_k max 200; page with search_after. Key role: reader.
find_paragraphs_advancedFind paragraphs advancedPOSTSame search as find_paragraphs in the JSON body form, which adds a filter_expression object, graph_query, query_image, a raw vector, rank_fusion and reranker objects. Key role: reader.

Search result, 2 operations

search_knowledge_boxSearch knowledge boxGETSearch returning separate result lists for documents, paragraphs, sentences and full text; the vendor recommends find_paragraphs for most uses. top_k max 200. Key role: reader.
search_knowledge_box_advancedSearch knowledge box advancedPOSTSame search as search_knowledge_box in the JSON body form, which adds a filter_expression object, sort, a raw vector and query_image. Key role: reader.

Suggestion, 2 operations

suggestSuggestGETAutocomplete: suggest paragraphs and entities that match a partial query, for type-ahead search. Key role: reader.
suggest_advancedSuggest advancedPOSTSame as suggest in the JSON body form, which accepts a filter_expression object and security groups. Key role: reader.

Agent run, 1 operation

run_resource_agentsRun resource agentsPOSTRun the Knowledge Box ingestion agents on one resource and return what they produced; agent_ids or filters restrict which agents run (all by default). Key role: reader.

Conversation field, 1 operation

set_conversation_fieldSet conversation fieldPUTAdd or replace a conversation field (a list of messages, each with a unique ident) on a resource. It is then processed and indexed. Key role: writer.

Counter, 1 operation

get_countersGet countersGETCount what the Knowledge Box holds: resources, paragraphs, fields and sentences, plus the index size. Key role: reader or manager.

Export, 1 operation

get_export_statusGet export statusGETRead the status of a Knowledge Box export started from the dashboard or the SDK: status, total, processed and retries. Key role: manager or reader.

Feedback, 1 operation

send_feedbackSend feedbackPOSTRecord whether an answer from ask_knowledge_box or ask_resource was good: pass its learning_id as ident, good true or false, task CHAT. Key role: reader.

Graph node, 1 operation

search_graph_nodesSearch graph nodesPOSTSearch the Knowledge Box knowledge graph and return matching nodes (entities). query is a graph node query object; top_k defaults to 50. Key role: reader.

Graph path, 1 operation

search_graph_pathsSearch graph pathsPOSTSearch the Knowledge Box knowledge graph and return paths (source node, relation, destination node). query is a graph query object; top_k defaults to 50. Key role: reader.

Graph relation, 1 operation

search_graph_relationsSearch graph relationsPOSTSearch the Knowledge Box knowledge graph and return matching relation types (edges). query is a graph relation query object; top_k defaults to 50. Key role: reader.

Import, 1 operation

get_import_statusGet import statusGETRead the status of a Knowledge Box import started from the dashboard or the SDK: status, total, processed and retries. Key role: manager or reader.

Key value field, 1 operation

set_key_value_fieldSet key value fieldPUTAdd or replace a key-value field on a resource; data must follow the key-value schema named by field_id (see list_kv_schemas). Key role: writer.

Knowledge box, 1 operation

get_knowledge_boxGet knowledge boxGETRead the Knowledge Box the API key belongs to: uuid, slug, title, description and hidden-resource and security settings. Key role: manager or reader.

Link field, 1 operation

set_link_fieldSet link fieldPUTAdd or replace a link field on a resource: the uri is fetched, its content extracted and indexed. Optional headers, cookies and localstorage are sent to that site. Key role: writer.

Processing status, 1 operation

get_processing_statusGet processing statusGETList resources by processing state (scheduled, completed, failed, timestamps), cursor paged: pass the returned cursor to get the next page. Key role: reader.

Resource paragraph, 1 operation

search_resourceSearch resourceGETSearch inside one resource and return matching paragraphs and sentences; query is required, top_k max 200. Key role: reader.

Summary, 1 operation

summarize_resourcesSummarize resourcesPOSTSummarize one or more resources (ids or slugs) with the Knowledge Box LLM; summary_kind simple (default) or extended. Returns a summary per resource and a consolidated one. Key role: reader.

Text field, 1 operation

set_text_fieldSet text fieldPUTAdd or replace a text field on a resource (body plus format such as PLAIN, MARKDOWN or HTML; at most 2 MB of text per request). It is then processed and indexed. Key role: writer.

Use Progress Agentic RAG from Claude Code, Cursor or Codex

Connect Progress Agentic RAG in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Progress Agentic RAG operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Progress Agentic RAG

Create your workspace
Sign up free. Nothing to install.
Connect Progress Agentic RAG once
Add your Progress Agentic RAG connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.