Integration

Valyu integration

Valyu is a search and research API for AI agents. Search the web, news, academic papers and licensed datasets (SEC filings, clinical trials, patents, market data) with ranked, full-text results; extract clean content, summaries or structured fields from URLs; get a cited AI answer; and run DeepResearch tasks and batches that return cited reports, PDFs and files, from instant answers to long investigations. Also lists the dataset catalogue and manages saved research workflows (beta). Every search, extraction, answer and research task spends credits from the account balance. Authentication is an API key from the Valyu platform, sent in the X-API-Key header.

Connect Valyu How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Valyu.Built from the 29 Valyu tools below, on the connection you add once.

What you can do with Valyu

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.

Deep research task, 8 operations

add_deep_research_instructionAdd deep research instructionPOSTAdds a follow-up instruction to a running DeepResearch task to steer it. Accepted only before the writing phase begins (400 otherwise).
cancel_deep_researchCancel deep researchPOSTCancels a DeepResearch task that is queued, running, awaiting_input or paused. Cancelling a running or awaiting_input task still charges the full mode price; a queued or paused one is free.
create_deep_researchCreate deep researchPOSTStarts a billed DeepResearch task that returns a cited report: the mode's price is reserved now (instant 8 credits ... max 2,500). Returns deepresearch_id at once; poll get_deep_research_status until completed.
delete_deep_researchDelete deep researchDELETEPermanently deletes a finished, failed or cancelled DeepResearch task with its report, sources, PDF and images. Cannot be undone; a queued or running task must be cancelled first.
get_deep_research_statusGet deep research statusGETReturns a DeepResearch task's status (queued, running, awaiting_input, paused, completed, failed, cancelled), progress and, once completed, the report in output with sources, pdf_url, images, deliverables and cost.
list_deep_research_tasksList deep research tasksGETLists the DeepResearch tasks of the API key's account: id, query, status, creation time and whether results are public. Read a task's report with get_deep_research_status.
respond_to_deep_researchRespond to deep researchPOSTAnswers a human-in-the-loop checkpoint of a DeepResearch task in awaiting_input or paused status, and resumes it. Read interaction_id and the checkpoint type from get_deep_research_status.
set_deep_research_publicSet deep research publicPOSTShares or unshares a DeepResearch task: public=true makes its results readable by anyone with the share link, without authentication; false makes them private again. Only the task owner can change it.

Research workflow, 6 operations

create_research_workflowCreate research workflowPOSTBeta. Creates a private research workflow (a DeepResearch template with typed variables) for your organization, within its quota (default 100). Running it later with create_deep_research is billed.
delete_research_workflowDelete research workflowDELETEBeta. Deletes one of your organization's research workflows (soft delete). Valyu-curated workflows cannot be deleted.
get_research_workflowGet research workflowGETBeta. Returns a research workflow with its variables, prompt template, strategy, report format, deliverables and recommended mode, at its current version or at version.
list_research_workflowsList research workflowsGETBeta. Lists research workflows (saved DeepResearch templates) visible to your organization: its own and Valyu-curated ones. Run one with create_deep_research, passing its slug as workflow_id.
preview_research_workflowPreview research workflowPOSTBeta. Resolves a research workflow's template with workflow_params without running it (free): returns the prompt, strategy, report format, deliverables, mode and tools a run would use, and the estimated credits.
update_research_workflowUpdate research workflowPATCHBeta. Edits one of your organization's research workflows: title and other metadata change in place; a version object publishes a new immutable version and needs a changelog. Valyu-curated workflows are read-only.

Research batch, 4 operations

cancel_research_batchCancel research batchPOSTCancels every queued and running task of an open or processing batch. Cannot be undone.
create_research_batchCreate research batchPOSTCreates an empty DeepResearch batch whose mode, output formats and search settings apply to every task; add tasks with add_research_batch_tasks. Creating it is free; each task added is billed.
get_research_batchGet research batchGETReturns a DeepResearch batch's status (open, processing, completed, completed_with_errors, cancelled), its task counts per status and its aggregated cost.
list_research_batchesList research batchesGETLists the DeepResearch batches of the API key's account with their status, mode, task counts and aggregated cost.

Datasource, 2 operations

list_datasourcesList datasourcesGETLists Valyu datasets with example queries, price per 1,000 results, required_plan and access_modes; a listing is not a promise of access, so check both before using an id in included_sources of search. Category filter: list_datasource_categories.
search_datasourcesSearch datasourcesGETFinds datasets from a plain-language description (for example FDA drug labels or CPI inflation), ranked best first with a relevance_score. Pass the returned ids as included_sources to search.

Research batch task, 2 operations

add_research_batch_tasksAdd research batch tasksPOSTAdds up to 100 billed research tasks to an open or processing batch; they inherit its mode. All their mode prices are reserved at once: if the balance cannot cover them all, 402 and none is added.
list_research_batch_tasksList research batch tasksGETLists the tasks of a DeepResearch batch, filterable by status and paged with last_key. include_output=true adds each task's report, sources, images, PDF link, deliverables and cost.

Research workflow version, 2 operations

get_research_workflow_versionGet research workflow versionGETBeta. Returns a research workflow resolved at one immutable version, in the same shape as get_research_workflow.
list_research_workflow_versionsList research workflow versionsGETBeta. Lists the immutable versions of a research workflow, newest first.

Answer, 1 operation

answer_questionAnswer questionPOSTSearches web, academic and financial sources and writes a cited AI answer; waits for the whole stream. The chunk carrying contents holds the full answer, search_results and cost. Billed for its searches, extraction and AI use.

Contents job, 1 operation

get_contents_jobGet contents jobGETReturns the status of an async extraction job from extract_contents (pending, processing, completed, partial, failed) and, once finished, one result per URL and the cost. Call again until the status is no longer pending or processing.

Datasource category, 1 operation

list_datasource_categoriesList datasource categoriesGETLists the dataset categories (such as research, healthcare, markets, company, economic, predictions, transportation, legal, politics, patents) used to filter list_datasources.

Page content, 1 operation

extract_contentsExtract contentsPOSTExtracts clean content from up to 10 URLs and returns it in results. With async=true (up to 50 URLs) it returns a job_id instead: poll get_contents_job. Billed 0.2 credit per extracted URL, plus 0.2 with a summary.

Search result, 1 operation

searchSearchPOSTSearches the web, news, papers and licensed datasets and returns ranked results with full extracted content. Billed per result by source (web 0.1 credit, SEC filings 1.92), so keep max_num_results low.

Use Valyu from Claude Code, Cursor or Codex

Connect Valyu in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Valyu 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 Valyu

Create your workspace
Sign up free. Nothing to install.
Connect Valyu once
Add your Valyu 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.