Provider REST API

16 actions

Connect Claude to Arize Phoenix

Read LLM traces, spans, sessions, datasets, experiments and prompts in Arize Phoenix, and annotate spans. Toolspoke puts 16 of its actions behind one MCP endpoint that Claude, Cursor, and Codex all speak.

Connect View pricing

Connection: Provider REST API
Authentication: API key
Actions exposed: 16
What a call bills: Nothing
Adapter: Maintained by Toolspoke

Connected in three steps

1
Install Arize Phoenix
Open the marketplace in your workspace, add Arize Phoenix to the project your agents work in, and it appears on the gateway immediately.
2
Connect the credential
Authenticate with api key. Where to get one, and what it has to be able to reach, is the next section.
3
Point your agent at the gateway
Give your client one address, https://toolspoke.com/mcp. Claude Code takes it as a command, Claude and Claude Desktop add it as a custom connector, and Cursor, Codex and VS Code each read it from a config file of their own.

.mcp.json

{
  "mcpServers": {
    "toolspoke": {
      "type": "http",
      "url": "https://toolspoke.com/mcp"
    }
  }
}

One block covers every tool you have installed. Arize Phoenix shows up in the client as soon as your policy allows it, and so does everything else you install later.

Where the address goes, per client

Claude Code: Run it in your project, then /mcp to sign in
claude mcp add --transport http toolspoke https://toolspoke.com/mcp
Claude and Claude Desktop: Settings, then Connectors, then Add custom connector
https://toolspoke.com/mcp
Cursor: ~/.cursor/mcp.json, or .cursor/mcp.json for one project
{ "mcpServers": { "toolspoke": { "url": "https://toolspoke.com/mcp" } } }
Codex: ~/.codex/config.toml
[mcp_servers.toolspoke] url = "https://toolspoke.com/mcp"
VS Code: .vscode/mcp.json, or the MCP: Add Server command
{ "servers": { "toolspoke": { "type": "http", "url": "https://toolspoke.com/mcp" } } }

What Arize Phoenix asks for

API key. You provide it once, when you install the connector. Toolspoke encrypts it at rest and decrypts it only for the length of a single call, and the gateway attaches it to the outbound request itself, so it is never part of the arguments an agent sends.

Phoenix URLRequired: Phoenix Cloud: the space URL from your browser's address bar, https://app.phoenix.arize.com/s/<space>. Self-hosted: the address your Phoenix serves its UI on, with no path. It must be reachable from the internet; a private-network address is refused.; https://app.phoenix.arize.com/s/your-space
API keyOptional: Phoenix → Settings → API Keys → System Keys (or your own user key) → Create key. Leave it empty only for a self-hosted Phoenix that runs without authentication.; eyJhbGciOi…

What Claude can do in Arize Phoenix

16 actions, each one declared and named by the connector rather than discovered at runtime. A workspace policy grants a person all of them, a hand-picked selection, everything on the read side, everything on the write side, or none.

Reads: 14Reads
Writes: 1Writes
Destructive: 1Destructive

Reads

Fetches data and changes nothing.

list_projects
List tracing projects by id and name. Every trace, span and session call takes a project id or name from here. Also the install's health check.
get_project
Fetch one project by id or name.
list_traces
List a project's traces, newest first by default, with latency and error counts. Narrow with filter, e.g. error_count > 0 or latency_ms >= 2000, or by session. include_spans returns every span with its inputs and outputs, which is large and holds your users' content.
list_spans
List a project's spans, filtered by trace, span kind (LLM, TOOL, RETRIEVER, AGENT…), status, name, parent or attribute. Pass parent_id "null" for root spans only. Spans carry inputs and outputs.
list_sessions
List a project's sessions (multi-turn conversations), newest first by default.
list_span_annotations
Read the annotations (evaluation labels, scores, human feedback) on some spans.
list_datasets
List datasets used for experiments and evaluations.
get_dataset
Fetch one dataset by id, with its example count and versions.
get_dataset_examples
Read a dataset's examples - inputs, expected outputs and metadata - at the latest or a named version, optionally one split.
list_experiments
List the experiments run against a dataset.
get_experiment
Fetch one experiment, with its run counts and the project its traces went to.
list_prompts
List the prompts stored in Phoenix.
get_latest_prompt
Fetch the latest version of a prompt - template, model and invocation parameters.
get_prompt_by_tag
Fetch the prompt version carrying a tag, e.g. production.

Writes

Creates or updates something on the other side.

annotate_span
Add or update one annotation on a span - a label, a score, or both, with an explanation. An annotation with the same name and identifier on the same span is replaced.

Destructive

Deletes or permanently alters something. Worth granting on purpose.

delete_trace
Delete one trace and all its spans and annotations permanently. There is no undo.

What it will not do

Enforced by the gateway rather than left to convention, which is why each of these can be stated flatly.

It cannot call anything else: The 16 actions above are the whole of it. A call to any other name is refused before it reaches Arize Phoenix rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time.
It reaches no further than your credential: Toolspoke holds no access to Arize Phoenix of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either.
It never hears from Arize Phoenix: Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Arize Phoenix. An agent has to ask.
It does not smooth over provider limits: Toolspoke does not retry, queue or back off around Arize Phoenix's own rate limits. A call that Arize Phoenix refuses comes back to the agent as a failed call.

Before you connect it

What can Claude do in Arize Phoenix?

16 named actions: 14 that only read, 1 that write and 1 that delete or permanently alter something. They include list_projects, get_project and list_traces. Nothing outside that list is reachable: the connector declares each operation by name rather than proxying whatever an agent asks for.

What credentials does the Arize Phoenix connector need?

API key. The connector asks for phoenix url, and optionally api key. Values are encrypted at rest and attached to the outbound request by the gateway, so they are never part of the arguments an agent sends and never reach the audit log.

Does the Arize Phoenix connector work with Cursor and Codex, or only Claude?

Any client that speaks MCP, and every one of them gets the same 16 actions. There is a single address, https://toolspoke.com/mcp. Claude Code adds it with claude mcp add --transport http, Claude and Claude Desktop take it as a custom connector in settings, Cursor reads it from .cursor/mcp.json, Codex from ~/.codex/config.toml, and VS Code from .vscode/mcp.json. Each of them signs in to the gateway itself, so there is no key to paste.

What does the Arize Phoenix connector not do?

The 16 actions above are the whole of it. A call to any other name is refused before it reaches Arize Phoenix rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time. Toolspoke holds no access to Arize Phoenix of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either. Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Arize Phoenix. An agent has to ask. Toolspoke does not retry, queue or back off around Arize Phoenix's own rate limits. A call that Arize Phoenix refuses comes back to the agent as a failed call.

Can I limit which actions an agent can call?

Yes, in two places. The project switches Arize Phoenix's actions on and off one at a time, for everyone in the project at once, and the screen groups them by read, write and destructive so turning off everything that deletes is one click. An individual agent key can then be narrowed further, to particular toolkits in a project and to particular actions in a toolkit. Whatever it was granted, a key never reaches a project its owner cannot.

What gets recorded when an agent calls Arize Phoenix?

Every attempt, with the agent that made it and the person that agent belongs to, the full request payload, the response payload, the status, the duration, and what the call cost in sandbox time and model tokens. Values whose key names a secret are masked out before the record is shown to anyone. An operation the connector marks as not retained never has its response body written at all, so the gateway keeps no second copy of what was read.

More observability connectors

Browse all of them