Pipecat PhoneLLM

# Fix list: Pipecat PhoneLLM

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/pipecat-phonellm, the October 2026 research run, assessed 17 October 2026. Grade C, 58.7 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Pipecat PhoneLLM: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Reliability, 53 out of 100, up to 9.4 more on the total

Why it scored 53: Hosted lines, read on Workers AI for Mini, the only per-token route. www.cloudflarestatus.com is a Statuspage with a Workers AI component (20). The incidents feed covers 7 to 16 October 2026 and holds one Workers AI incident, minor, on 15 October from 07:10 to 09:37 UTC, naming five other models and not PhoneLLM. The history page loads by script, so the rest of the 90 days went unread, and the model is two days old (10 of 30, as the Workers AI listing scored the same record). The limits page gives Text Generation 300 requests a minute, and the model page doesn't mark Mini as paid-only, though PhoneLLM isn't named (13 of 15). A 429 carries code 3036 for a spent free allocation or 3040 for capacity, and `rejectIfBusy` fails fast. No `Retry-After` is documented for 3040. Inference has no side effects, so a retry is safe (10 of 15). No SLA for Workers AI found (0). Workers AI is generally available and the model page has no beta label, but Daily names the model Alpha 2 and the limits page says beta models may get lower limits, so the surface isn't counted as GA (0). Max has no serverless route to read. 53.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 2. Payments & pricing, 40 out of 100, up to 7.5 more on the total

Why it scored 40: Read on Workers AI for Mini. Machine Payments takes x402 on `POST /ai/run` for four models, and PhoneLLM Mini isn't one of them (0). $0.06 per million input and $0.24 per million output tokens, published without a login on the model and pricing pages (20). 10,000 neurons a day are free on the Workers Free plan, about 1.8 million input tokens. We didn't sign up to confirm that no card is asked for, and scored it as the Workers AI listing did (20). A person signs up with Cloudflare in a browser for a token. The weights download from Hugging Face without an account, but other open-weight listings graded on a hosted route get no onboarding credit for that, so this one doesn't either (0). Max has no per-token price anywhere, only GPU time at Modal or Baseten. 40.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 3. Agent ergonomics, 61 out of 100, up to 6.3 more on the total

Why it scored 61: Model reading (tool use, structured output, caching, context, batch, SDKs, errors and defaults). Trained for multi-turn tool calls and Pipecat Flows, with OpenAI-style `tools`, `tool_choice` and `parallel_tool_calls` on Workers AI. Daily's PhoneBench gives 5.9% tool-call failures for Mini and 3.1% for Max (vendor-reported). Alpha 1 discussion 3 reports the model confirming details back in place of calling a tool, unanswered, and we don't know whether Alpha 2 fixes it (15 of 20). `json_object` and `json_schema` in the Workers AI schema. The Workers AI docs say adherence isn't guaranteed, and neither card mentions structured output (8 of 15). The vLLM recipes turn on prefix caching, but Alpha 1 discussion 1 reports it as a no-op for nemotron_h in vLLM, unanswered. Workers AI lists no cached price for Mini (5 of 15). 262,140 tokens on Workers AI and 262,144 in Mini's config, 1,048,576 in Max's config. The cards' recipes cap at 32,768 (13 of 15). The Mini model page has no Batch tag, so the Workers AI batch route isn't open to it. Self-hosted, the cards suggest 128 to 256 requests in flight per GPU (3 of 10). Any OpenAI SDK against `/v1/chat/completions`, the Workers binding, Cloudflare's TypeScript, Python and Go SDKs, and Pipecat's OpenAI-compatible services (9 of 10). Workers AI errors carry internal codes. Defaults work against the caller. Both repositories' generation_config.json sample by default (Max at temperature 1.0) where Daily recommends 0, and Max needs `chat_template_kwargs` on every request or its reply lands in `reasoning` (8 of 15). 61.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 4. Security & auth, 66 out of 100, up to 6 more on the total

Why it scored 66: Model reading. On Workers AI, Cloudflare API tokens limited to Workers AI permissions on chosen accounts, revocable, with expiry and IP filters, per the Workers AI listing's check of 9 October. Permission is per account, not per model (27 of 30). Cloudflare's data usage page and service terms say customer content isn't used to train models without consent, with clause 9 allowing use to provide and improve services, per the same check. Self-hosted, prompts stay on the owner's hardware (17 of 20). The model is built to take a caller's transcribed speech and call tools. Neither card nor the post has limitations, a safety evaluation, or guidance on injected or hostile input. The template wraps tool results in `<tool_response>` and does nothing to neutralise caller text. Mini self-hosted runs modelling code from the repository with `--trust-remote-code` (3 of 15). Direct Workers AI calls have no per-request log that we found. AI Gateway logs prompts and responses when a call is routed through it (6 of 15). Cloudflare's security.txt names a HackerOne programme, per the Workers AI check. Daily has no security.txt and a disclosure policy with no bounty, per the Pipecat check of 17 October, and the cards give no security contact (13 of 20). 66.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 5. Schema & documentation, 72 out of 100, up to 4.6 more on the total

Why it scored 72: Model reading on the API lines. Each Workers AI model page links JSON Schemas for input and output. Mini's synchronous input schema lists `required: ["prompt"]` and has no `messages` property, while the page's TypeScript, Python and curl examples and Cloudflare's changelog all send `messages`. It also carries `audio`, `modalities` and `web_search_options`, which a text-only model has no use for (18 of 25). developers.cloudflare.com/llms.txt and a Markdown copy of the model page and changelog. The cards are Markdown on Hugging Face (10). The post and cards say what each size is for (Mini for speed, Max for accuracy, both for phone agents with tool calling and Pipecat Flows). Neither card has a limitations or out-of-scope section, and English-only is in the metadata, not the text (12 of 20). `temperature` bounded 0 to 2, `tool_choice` an enum or a forced function, `response_format` a typed one-of, and `chat_template_kwargs.enable_thinking` typed and described (12 of 15). Examples in TypeScript, Python, curl and vLLM recipes for both sizes. The Workers AI examples omit the `temperature: 0` and `enable_thinking: false` that Cloudflare's changelog asks for, and the error table is Workers AI's generic one (11 of 15). The version is in the model ID (`alpha-2`), Alpha 1 stays published, and Hugging Face keeps commit history (three commits each). No changelog of PhoneLLM's own. Changes are listed in the post and Cloudflare's changelog entry (9 of 15). 72.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 6. Transparency & trust, 56 out of 100, up to 3.9 more on the total

Made of editorial 55, provenance 57.

Why it scored 56: Mini is BSD-2-Clause, an OSI licence, with NVIDIA's Nemotron Open Model License kept for the base work, and the card explains how the two apply. Max is OpenMDW-1.1, permissive but not on the OSI list. Training data and code aren't published, and the hosted route runs under Cloudflare's terms (22 of 30). The post says training data is fully synthetic, from unnamed open-weight teachers. On Workers AI, Cloudflare's no-training statements agree with each other apart from clause 9, per the Workers AI check. Self-hosted, nothing leaves the owner's hardware. Daily's privacy policy and DPA weren't re-read for the models (18 of 30). No deprecation policy or dated notices from Daily. Workers AI posts dated notices but has no policy (5 of 20). Cloudflare publishes a sub-processor page, and where a Workers AI request runs isn't stated beyond its network. The weights are safetensors, but Mini's repository ships Python modelling code that a self-hosted run executes (10 of 20). 55.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: pipecat.ai, registered 2024-05-01 (2 years) (7 of 15)
- Endpoint on the vendor's domain: api.cloudflare.com is not on pipecat.ai (0 of 15)
- Changelog: not found (0 of 10)
- security.txt: not found (0 of 10)

## 7. Maintenance & community, 58 out of 100, up to 3.7 more on the total

Why it scored 58: Model reading, with notice periods and churn in place of release counts. Alpha 2 released on 15 October 2026 (30). Two releases since 27 August, Alpha 1 and Alpha 2, with Alpha 1 still published. Daily states no notice period, support window or deprecation plan, the alpha label signals more changes, and Workers AI has given 13 to 22 days' notice before retiring models, per the Workers AI check (8 of 20). Three Alpha 1 discussions opened on 31 August and 16 September, a vLLM caching bug, a request for FP8 weights and a tool-call fault, have no reply from the maintainers on 17 October, though FP8 weights did ship with Alpha 2. Daily's Discord is linked from the post and wasn't sampled (4 of 25). vLLM 0.30 recipes measured by Daily, native Transformers support for Max from 5.8, and Workers AI with Cloudflare's SDKs. Mini still needs repository code (12 of 15). Max's generation config names Transformers 5.8.1 and Mini's 4.57.6. No CI or tests for the weights visible to us (4 of 10). 58.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- unchecked: Workers AI incident history before 7 October 2026, since the status history page loads by script and the feed starts there
- unchecked: Baseten's PhoneLLM deployment, which sits in Baseten's logged-in app, and the AWS SageMaker route, for which the post links only SageMaker's home page
- unchecked: whether Workers AI signup asks for a card. Scored as the Workers AI listing scored it
- unchecked: Daily's terms, privacy policy and security.txt for the model pages, taken from the Pipecat listing's check of 17 October 2026
- unchecked: the cards of the FP8 and NVFP4 repositories, and the PhoneBench harness and method, which Daily says will be published later
- Whether Alpha 2 fixes the Alpha 1 reports of tool-call confirmation in place of a call and of vLLM prefix caching having no effect on nemotron_h
- Whether Workers AI's input schema, which requires `prompt` and omits `messages`, is an error in the published schema or reflects what the endpoint validates. We did not call it
- firstReleased is 27 August 2026, when Alpha 1's card and licence were added and PhoneBench Alpha 1 is dated. The Alpha 1 repository was created on 24 August, and its history before 27 August was squashed, so whether it was public earlier is unknown

## Weaknesses

- Labelled Alpha 2 by Daily, with no deprecation policy, support period or changelog of its own
- Max has no per-token hosted route, only dedicated GPU endpoints or two B200 or B300 GPUs of the owner's
- No limitations section, safety evaluation or prompt-injection guidance, for a model that acts on a caller's speech
- Workers AI's published input schema requires `prompt` and has no `messages` property, while every example sends `messages`
- Three Alpha 1 discussions, one on tool calls and one on vLLM prefix caching, have no reply from the maintainers

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Send `temperature: 0` on every call. The repositories' generation_config.json samples by default, and Max's sets temperature 1.0
- For Max, serve with `--reasoning-parser nemotron_v3` and send `chat_template_kwargs` with `enable_thinking: true` on every request, or the reply lands in `reasoning`
- For Mini, serve without a reasoning parser and keep `enable_thinking` false, as Cloudflare and the card say, or tool calls can be filed under `reasoning`
- Confirm bookings, payments and transfers in your own handler before acting on a tool call, since caller speech reaches the model directly
- Pin the model ID including `alpha-2`. Daily states no notice period, and a later alpha may replace it

## What the review panel asked for

- a stated support window
- Add a cached-input price
- Give Max a per-token route
- Correct the published input schema
- Add a limitations section
- Publish a deprecation policy
- Reply to the Alpha 1 reports

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.