Integration

Helicone integration

Helicone LLM observability API. Query the requests an application sent to model providers with their prompts, completions, latency, token counts and cost, count them, read one by id, attach scores, thumbs feedback and custom properties to them, browse and tag sessions, aggregate per end user metrics, list the custom properties in use, and read the versioned prompts stored in Helicone. Reads are POST calls carrying a filter object, and every response is wrapped in a data and error envelope where error stays null on success.

Connect Helicone How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Helicone.Built from the 18 Helicone tools below, on the connection you add once.

Common Helicone use cases

marking a conversation as reviewed
loading the live prompt at application start
reporting weekly active users of an assistant
exporting yesterday's model calls for analysis
listing the prompts an organization has stored
checking the tags of a prompt before editing it
inspecting the exact prompt behind a bad answer
auditing which keys can write to the organization
spot checking the spend of a handful of named users
checking whether a conversation was already reviewed

What you can do with Helicone

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.

Request, 3 operations

count_requestsCount requestsPOSTReturns how many logged requests match a filter, without returning them.
get_requestGet requestGETReturns a single logged request with its prompt, completion and cost.
query_requestsQuery requestsPOSTReturns logged LLM requests with their prompts, completions, latency, tokens and cost.

Prompt, 2 operations

get_promptGet promptGETReturns a stored prompt with its name and tags.
query_promptsQuery promptsPOSTLists the prompts stored in Helicone with their tags.

Property, 2 operations

list_propertiesList propertiesPOSTReturns the distinct custom property names seen on logged requests.
set_request_propertySet request propertyPUTAdds or replaces one custom property on an already logged request.

Session tag, 2 operations

get_session_tagGet session tagGETReturns the tag stored on a session.
tag_sessionTag sessionPOSTSets a single tag on a session.

Api key, 1 operation

list_api_keysList api keysGETLists the organization API keys with their permissions and creation dates.

Feedback, 1 operation

add_request_feedbackAdd request feedbackPOSTRecords a thumbs up or thumbs down on a logged request.

Prompt version, 1 operation

get_production_prompt_versionGet production prompt versionPOSTReturns the version of a prompt currently promoted to production.

Score, 1 operation

add_request_scoreAdd request scorePOSTAttaches one or more named evaluation scores to a logged request.

Session, 1 operation

query_sessionsQuery sessionsPOSTReturns sessions with their request count, token totals, cost and average latency.

Session metric, 1 operation

query_session_metricsQuery session metricsPOSTReturns histograms and averages of session count, duration and cost.

Session name, 1 operation

query_session_namesQuery session namesPOSTReturns the session names matching a text fragment, each with how often and how recently it was used.

User, 1 operation

query_usersQuery usersPOSTReturns request counts, token totals and cost grouped by end user.

User metric, 1 operation

query_user_metricsQuery user metricsPOSTReturns activity metrics per end user, including first and last activity and averages.

Use Helicone from Claude Code, Cursor or Codex

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

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

Helicone integration FAQ

What can you do with the Helicone integration?

LiveContext exposes 18 operations on Helicone, covering request, prompt, property and session tag. Common uses include marking a conversation as reviewed, loading the live prompt at application start and reporting weekly active users of an assistant.