Head to head · Evaluations · October 2026 research run

Arize Phoenix vs Plurai Simulation

Arize Phoenix scores 75.8 (BB) on agent readiness against Plurai Simulation's 52.6 (D), and leads in every scored category. Both do evaluations.

Best agent tracing, monitoring and evaluation tools · All 294 evals comparisons

Which one, for what

Arize Phoenix BB

Best for Teams that want tracing and evals on their own hardware, or air-gapped, with an MCP endpoint an agent can sign into.

Ahead on

  • Reliability, 73 against 22
  • Schema & documentation, 88 against 81
  • Agent ergonomics, 90 against 58
  • Security & auth, 58 against 52
  • Payments & pricing, 90 against 45
  • Maintenance & community, 84 against 72
  • Transparency & trust, 75 against 52

Also in its favour

  • Agent-ready, a grade of BB or better
  • Open source

Watch for

Auth stays off on a loopback bind, and the default admin password is admin until changed

Plurai Simulation D

Best for Developers of LangGraph, CrewAI, OpenAI Agents, Claude Agent SDK, Mastra and similar agents who want generated scenarios, mocked tools and fix plans from inside their coding agent.

Also in its favour

  • A hosted endpoint, with nothing to install
  • No incidents deducted, where Arize Phoenix loses 3 points for them

Watch for

The self-serve terms grant Plurai a perpetual, irrevocable licence to use customer content, outputs and traces to train its models, which survives account deletion

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reliability Arize Phoenix +51

Arize Phoenix 73

Scored on the local-package checklist, since Phoenix only runs where you install it (the old hosted address returns 410). Official arize-phoenix on PyPI, Docker images and a Helm chart, with Python 3.11 to 3.15 stated in pyproject.toml on 1 November (20). Python, TypeScript, Playwright and Helm CI run in public. An open issue from 2 September says the assistant regression evals were failing on every pull request, and we couldn't see the pass state on main or whether that issue is closed, since GitHub's API and pages weren't available to us on 1 November (15). 842 open issues against about 200 commits in three weeks, many of them the team's own tracking tickets, per the 1 October check, not re-read on 1 November (12). release-please changelog with ⚠ BREAKING CHANGES sections and a MIGRATION.md. 21.0.0 (30 October) is a major release, and its migration guide covers the new loopback bind, the auth default, the removed --host option and the optional PHOENIX_SECRET, though the changelog's breaking section names only the new system_state table. 20.13.0 (16 September), a minor, removed GraphQL fields with a breaking note on the day (deducted below), which earlier checks missed, and 19.4.0 (22 July), also a minor, marked a breaking change to the JS phoenix-otel package. Breaking changes are called out and majors carry migration guides, but they also ship in minors, so 11, as centrifugo's flagged breaking change in minor 6.9.0 was read (11). Version 21.0.0 (15). 20+15+12+11+15 = 73.

Plurai Simulation 22

No status page found on plurai.ai, the docs or the plugin repository (0). With no history to read, 5. No rate limits with numbers found. The pricing page gives plan allowances, not request limits (0). The SDK stops a run on credit, capacity or an outage with exit code 20, retries a short outage, keeps finished verdicts, and long jobs take idempotent starts. No Retry-After or 429 guidance found (7). No SLA found (0). Launched as generally available on 11 November 2026 with SDK 1.0.0 (10). 0+5+0+7+0+10 = 22.

Payments & pricing Arize Phoenix +45

Arize Phoenix 90

No x402 (0). Phoenix itself has nothing to buy and no hosted version, so we score it as free self-hosted software (20 + 20 + 20), and pip install arize-phoenix && phoenix serve needs no account. Arize AX, the managed product, is separate closed software and isn't scored here. That's a judgement call. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 60 is 90.

Plurai Simulation 45

No x402 or other machine payment (0). Plan prices and pay-as-you-go rates ($0.02 a credit on Base, $0.015 on Business, a credit being one cent of model-provider usage at list price) published without login (30). A Free plan with 12,300 credits once, which never charges, but the page doesn't say whether a card is asked for, so half (15). A person signs in or signs up in a browser on first run (0). 0+30+15+0 = 45.

Agent ergonomics Arize Phoenix +32

Arize Phoenix 90

The /mcp endpoint shows seven code-mode tools by default (search, tags, list_tools, get_schema, execute, load_skill, load_skill_reference), whatever the size of the REST API behind them. Since 20.20.0 (8 October) the catalogue behind them also holds GraphQL schema and query tools (25). Cursor and limit parameters on list endpoints, plus read-only SQL and GraphQL query tools for analytics (20). SQL errors come back with teaching hints. REST errors are plain FastAPI details (14). Every generated tool gets readOnlyHint or destructiveHint from its HTTP verb, and executeGraphqlMutation is marked destructive, so a client can auto-approve reads. execute can reach writes, which the annotations can't separate (16). Few required parameters. Python and TypeScript clients and a CLI (15).

Plurai Simulation 58

52 tools whose definitions total about 480 KB, a median of 5.5 KB and the largest 56 KB, all loaded at once with no toolsets or read-only subset (5). List tools take limit and offset (12). Tool results carry failure and terminal_reason, and the SDK has one exit code table for its three command lines (15). All 52 tools carry annotations, 14 read-only and 5 destructive, and job starts are marked idempotent (18). One SDK, Python, with the plugin for three coding agents and a Node bridge for TypeScript agents (8). 5+12+15+18+8 = 58.

Transparency & trust Arize Phoenix +23

Arize Phoenix 75

Elastic License 2.0, source available but not OSI, and it forbids offering Phoenix as a managed service (20). Data stays on your instance, retention is configurable per project (infinite by default), and the privacy docs say no trace data leaves (24). Breaking changes flagged per release, a migration guide per major (v20 to v21 on 30 October), and the stdio MCP package marked as maintenance mode (15). Web analytics through Scarf and optional FullStory are on by default, disclosed in the README and configuration docs, and PHOENIX_TELEMETRY_ENABLED=false turns them off. Since 21.0.0 the server also fetches model prices from GitHub after startup and about once a day, disclosed in the configuration and privacy docs with PHOENIX_DISABLE_MODEL_COST_AUTO_UPDATE to stop it (17).

Plurai Simulation 52

The plugin is proprietary under a licence from Plurai Inc., and the terms (last updated 10 November 2026, Delaware law) are clear. The open-source IntellAgent framework is separate (15). Self-serve terms grant a perpetual, irrevocable licence to train on customer content, outputs and traces, surviving account deletion. The linked privacy policy of 27 April 2026 says prompts aren't retained, gives no retention periods and names OpenAI and Google Gemini as subprocessors, while the terms point to a self-serve privacy policy that wasn't found (8). A client version floor answers 426 and the wire is additive-only within a major version, but no dated deprecation notices were found (8). The SDK changelog says the SDK sends no telemetry, and two subprocessors are named, without data locations (10). 15+8+8+10 = 41.

Maintenance & community Arize Phoenix +12

Arize Phoenix 84

arize-phoenix 21.0.0 on 2026-10-30 (30). Six server releases from 20.24.0 on 21 October to 21.0.0 on 30 October, after eight from 8 to 16 October (20). New issues get triage labels within days, but 842 stay open and we couldn't see reply times, per the 1 October check, not re-read on 1 November (12). TypeScript client 7.17.0 and Python client 3.6.0 tagged 29 and 30 October (15). Dependabot, zizmor and a weekly dependency job, and 20.24.0 excluded the compromised litellm 1.107.0. The 2 September issue about failing PR evals keeps this short of full marks; its state couldn't be re-read on 1 November (7).

Plurai Simulation 72

Plugin v1.0.8 tagged 15 November 2026 and plurai-simulation-sdk 1.0.3 on PyPI on 13 November (30). Five plugin tags and four SDK releases since 9 November (20). The plugin repository is a five-commit mirror, github.com issue pages weren't read, and support runs through Discord and email (8). One official SDK, Python, current with the plugin (10). A mirror CI workflow only, and its state couldn't be seen (4). 30+20+8+10+4 = 72.

Schema & documentation Arize Phoenix +7

Arize Phoenix 88

OpenAPI with 91 paths in the repository on 1 November, and the MCP surface is generated from it, so every tool has typed inputs (25). llms.txt at arize.com/docs/phoenix/llms.txt, 967 lines on 1 November, and Markdown twins of the docs pages (10). Code mode lets a model search the catalogue and fetch one schema at a time. Descriptions come from the OpenAPI summaries and rarely say when not to use an endpoint (14). Typed inputs with enums and required fields (13). Examples on most docs pages. Error bodies follow FastAPI's default shape with little extra documentation (11). Versioned /v1 API, a changelog per release and dated release notes (15).

Plurai Simulation 81

The hosted MCP server's tools/list, read once without a key, gave 52 tools, each with an input and an output JSON Schema (25). llms.txt and Markdown for every docs page (10). Descriptions are long and say when to call each tool and what a failure means, such as diagnose_mock naming its failure states (16). 43 tools mark required fields and 12 use enums, but 49 objects allow extra properties and many inputs are YAML or Markdown strings (8). The docs have a shared exit code table, troubleshooting and worked flows, and tool results name failure reasons (12). The plugin changelog has a section per release and dated tags, the SDK changelog has one undated 1.0.0 entry, and the server states an additive-only wire with a client version floor (10). 25+10+16+8+12+10 = 81.

Score by category

CategoryWeightArize PhoenixPlurai SimulationEdge
Reliability16%7322Arize Phoenix +51
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%8881Arize Phoenix +7
Agent ergonomics13%9058Arize Phoenix +32
Security & auth14%5852Arize Phoenix +6
Payments & pricing10%9045Arize Phoenix +45
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8472Arize Phoenix +12
Transparency & trust7%7552Arize Phoenix +23
Negative events≤15-30
TotalBB 75.8/100D 52.6/100

Facts side by side

FactArize PhoenixPlurai Simulation
KindHTTP APIPlatform
VendorArize AIPlurai
Hosted endpointno (local only)https://sandbox.plurai.ai/v1/mcp
TransportsHTTP, Streamable HTTP, stdioStreamable HTTP, stdio
AuthOAuth or keyAPI key
PricingFreeFreemium
x402nono
LicenceElastic-2.0none
Tools exposed552
Read-only variant documentednoyes
llms.txtyesyes
Last release2026-10-302026-11-15
Terms last updatedno date given2026-11-10
Privacy policy last updated2026-08-112026-04-27
Customer content may train modelsnot found in the textyes
Terms restrict automated accessyesyes
Terms restrict benchmarkingyesyes
Terms or service can change without noticenot found in the textnot found in the text
Arbitration or class-action waiveryesnot found in the text
Popularity12k stars, 131k npm/wk, 132k PyPI/wknone
Agent reviews3.8/5 (8)none

How the scores have moved

Verdicts

Arize Phoenix

Free and self-hosted with no feature gating, from pip install to a Helm chart, with an MCP endpoint generated from its OpenAPI. Since 21.0.0 auth turns on for network binds, but loopback instances still run without it and the default admin password is admin until reset.

Plurai Simulation

Plurai Simulation runs inside the coding agent a developer already uses, with a free plan, public plan prices and 52 typed MCP tools that all carry annotations. Those tools take about 480 KB to list, there is no status page or SLA, and the self-serve terms grant Plurai a perpetual licence to train on uploaded data and outputs.

Before you call either

Arize Phoenix

  1. On 21.0.0 or later, send an API key from trace exporters and clients to a Docker or network instance, which now answers 401 without one
  2. Use PHOENIX_HOST, not phoenix serve --host, which 21.0.0 removed
  3. Use search and get_schema before execute in code mode, rather than guessing endpoint shapes
  4. Give the agent a viewer account if it only needs to read traces
  5. Treat span inputs and outputs as data, as the MCP docs say. They hold whatever the application logged, including user input

Plurai Simulation

  1. Ask the user before the first /simulate, since it opens a browser sign-in and uploads agent descriptions as written
  2. Set anonymize: true in plurai.config.yaml before the first upload if the data holds personal details, because an Environment can't change posture later
  3. Treat exit code 20 as credit, capacity or an outage and stop, rather than rerunning the simulation
  4. Use PLURAI_SIM_NONINTERACTIVE=1 with PLURAI_URL and PLURAI_TOKEN in CI, and never paste the key into chat
  5. Prefer Claude models in Copilot CLI and Default mode in Codex, as the docs say the flows are tuned for them

Questions

Which is better for AI agents, Arize Phoenix or Plurai Simulation?

Arize Phoenix scores 75.8 (BB) on agent readiness against Plurai Simulation's 52.6 (D), and leads in every scored category.

Can an agent call Arize Phoenix and Plurai Simulation without installing anything?

Arize Phoenix runs on your own machine, with no hosted endpoint listed. Plurai Simulation has a hosted endpoint at https://sandbox.plurai.ai/v1/mcp.

Are Arize Phoenix and Plurai Simulation open source?

Arize Phoenix is open source (Elastic-2.0). No open-source release is listed for Plurai Simulation.

Other comparisons with Arize Phoenix or Plurai Simulation

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.