# Inworld Realtime API vs Rime Prism (slim)

> Inworld Realtime and Rime Prism score within a point of each other for voice agents, 60.1 and 59.7 out of 100. Prices, MCP, x402, uptime and agent notes side by side.

- Full: https://www.anchorterminal.com/compare/inworld-realtime-vs-rime-prism.md (~5,300 tokens) · this version ~2,680 tokens · JSON https://www.anchorterminal.com/compare/inworld-realtime-vs-rime-prism.json · canonical https://www.anchorterminal.com/compare/inworld-realtime-vs-rime-prism
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-11-07
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)

Inworld Realtime API and Rime Prism score within a point of each other on agent readiness, 60.1 (C) and 59.7 (C). Rime Prism leads on schema & documentation, agent ergonomics and transparency & trust. Both do voice agents.

- Inworld Realtime API: C 60.1, rank #668 of 1486 · https://www.anchorterminal.com/tools/inworld-realtime.min.md
- Rime Prism: C 59.7, rank #685 of 1486 · https://www.anchorterminal.com/tools/rime-prism.min.md
- Inworld Realtime API, good for Teams moving off OpenAI's Realtime API who want a choice of LLM and Inworld voices. Ahead on Reliability, 62 against 55.
- Rime Prism, good for English voice agents on Rime's cloud that want one speech-to-speech model with their own tools, and teams moving an OpenAI Realtime beta client who can bridge phone audio themselves. Ahead on Schema & documentation, 69 against 60; Agent ergonomics, 73 against 68; Transparency & trust, 72 against 63. Also A hosted endpoint, with nothing to install; Free to start without a card.

| Category | Inworld Realtime API | Rime Prism |
| --- | --- | --- |
| Reliability (16%) | 62 | 55 |
| Performance (10%) | pending | pending |
| Schema & documentation (13%) | 60 | 69 |
| Agent ergonomics (13%) | 68 | 73 |
| Security & auth (14%) | 58 | 56 |
| Payments & pricing (10%) | 40 | 40 |
| Task success (10%) | pending | pending |
| Maintenance & community (7%) | 71 | 75 |
| Transparency & trust (7%) | 63 | 72 |

| Fact (where they differ) | Inworld Realtime API | Rime Prism |
| --- | --- | --- |
| Vendor | Inworld AI | Rime Labs |
| Hosted endpoint | no (local only) | `wss://prism.api.rime.ai/v1/realtime` |
| Transports | websocket, HTTP | websocket |
| Pricing | Pay per use | Freemium |
| Price for voice agents | $25 per 1M characters | not published |
| Licence | none | Proprietary hosted service under Rime's terms. The Rime SDK used to call it (github.com/rimelabs/rime-sdk) is MIT. |
| Tools exposed | 0 | none |
| Last release | 2026-10-16 | 2026-11-01 |
| Terms last updated | 2025-06-11 | no date given |
| Privacy policy last updated | 2026-01-18 | no date given |
| Terms restrict benchmarking | not found in the text | yes |
| Arbitration or class-action waiver | not found in the text | yes |

Why the edge:
- Schema & documentation, Rime Prism +9: Inworld Realtime API 60 (No AsyncAPI or other machine-readable file for the socket, and the spec paths redirect (0). llms.txt, Markdown twins and a docs MCP server (10). The overview, OpenAI migration guide, `providerData` reference and the billing page's account of how a session is counted say what to use when (16 of 20). The reference lists every client and server event with field types and required marks, as text rather than a schema (12 of 15). JavaScript and Python example projects and an error event with `type`, `code` and `param`, without a list of codes (9 of 15). Dated release notes with affected clients and required actions, but no API version beyond `/v1` (13 of 15). Total 60.); Rime Prism 69 (No OpenAPI or AsyncAPI file for the WebSocket API was found in llms.txt or the reference (0 of 25). llms.txt and a Markdown twin of every page (10). The overview says what Prism handles, what it doesn't (text-only replies, sampling settings, a choice of language model, ending the caller's turn) and when a pipeline fits better, and a chart says when to use the SDK, LiveKit, Pipecat or the raw WebSocket (18 of 20). The WebSocket reference types every client and server event field with defaults and allowed values, refuses unknown fields with `invalid_client_event` naming the field, and takes tool parameters as JSON Schema. The schemas exist only as tables, not a file (13 of 15). Python and Node.js examples on each guide, and an errors page that lists 41 codes with cause, scope and what to do, plus HTTP refusals and close codes (15). A dated changelog with an RSS feed, a `version` field in `session.created` (such as `20261023a`) and pinned SDK versions. The path carries `/v1`, with no version policy for the event set found. Partly met, so the earlier 13 is kept (13 of 15). Docs and changelog re-read on 4 November. 0+10+18+13+15+13 = 69.)
- Transparency & trust, Rime Prism +9: Inworld Realtime API 63 (Closed service with clear terms naming Theai, Inc. dba Inworld AI (15). The ZDR page says what a session keeps and that third-party LLM requests aren't covered, but the terms and privacy notice disagree on training (12 of 30). Release notes flag potentially breaking changes with the clients affected and what to do, without stating a notice period (14 of 20). US headquarters with EU and India residency on Enterprise, and no sub-processor list in the docs, while sessions may route to third-party LLM providers (10 of 20). Total 51.); Rime Prism 72 (Closed service with clear terms from Rime Labs, Inc., and the SDK is MIT (15 of 30). The terms changed on 29 October 2026. Section 6.3(b) now grants Rime a perpetual licence to train and improve its models on Free Plan customer data (excluding voice recordings, audio with a natural person's voice and third parties' personal data), while the privacy policy (effective 22 August 2025) says "We do not use customer data to train AI models unless explicitly requested and agreed upon" and the security page says "We never use customer text or audio to train our models". The privacy and security pages still describe text in and audio out, and none covers Prism's caller audio, while the new terms define Call Content and say not to send Caller audio on a Free Plan. The DPA at rime.ai/dpa couldn't be reached from our network on 28 October. The contradiction between the terms and the security page is charged once, as the misleading-claim deduction, so the line holds (15 of 30). The new terms promise at least 30 days' notice before retiring a Model Version used on a paid plan. Arcana's cloud retirement came with 17 days' notice before that policy (15 of 20, was 10). A sub-processor list with locations, all in the United States, and the Prism endpoint stated as US East (us-east-2), per the 28 October check (18 of 20). 15+15+15+18 = 63.)
- Reliability, Inworld Realtime API +7: Inworld Realtime API 62 (Hosted reading of the checklist. status.inworld.ai is a public status page (20), with history we couldn't read, as on inworld-tts (5). Concurrent sessions are published per plan, 10 on On-Demand up to 1,000 on Growth (15). The rate limits page gives exponential backoff with jitter and Retry-After on longer windows, and sessions emit `rate_limits.updated`, with nothing on reconnecting a dropped session (12 of 15). An SLA appears only in the Enterprise column, with no terms (0). No preview or beta label on the Realtime API (10). 20 + 5 + 15 + 12 + 0 + 10 = 62.); Rime Prism 55 (status.rime.ai redirects to an incident.io page at rimestat.us with Text-to-speech API and Platform API components, and still no Prism component on 4 November 2026. A vendor status page with no component for the product takes 15 in this run (15, was 10). Its feed lists two incidents, Text-to-speech partial degradations on 3 and 5 September, and nothing since Prism launched on 28 October, a short record of seven days (15, was 5). The errors page publishes per-session numbers (55 seconds a turn, 60 seconds of queued speech, 120 seconds or 512 tokens a reply, 5 minutes idle, 1 MiB a message), but no concurrency limit per account or key, and a full server answers 503. Partly met, so the earlier 8 is kept (8). Overload is documented as `capacity_exceeded`, `unavailable`, a 503 refusal ("Retry after a backoff") and close code 1001, the page says which reply failures can succeed when sent again, and the tools guide says never to rerun a call with an unknown outcome. Safe-retry guidance without idempotency keys gives 12 in this run (12, was 13, not a checklist value). SLAs are an Enterprise item and none is published (0). The changelog calls Prism available, but all four SDKs are 0.1.0 alpha releases, the Pipecat path uses a pinned preview, and the pricing page says half duplex at launch. Partly met, so the earlier 5 is kept (5). 15+15+8+12+0+5 = 55.)
- Agent ergonomics, Rime Prism +5: Inworld Realtime API 68 (Scored as an API. Conversation history defaults to the last 100 items, can be trimmed, and can be cached on Anthropic models, and every `response.done` carries a `usage` block (18 of 25). No lists to page. Output controls are modalities, turn detection settings, optional transcription and `response.speak` (12 of 20). Error events follow OpenAI's `type` and `code` shape, unsupported fields come back as non-fatal `unsupported_field` warnings, and `item_not_found` names the bad parameter (14 of 20). Tools run on the client and a follow-up response starts automatically after a tool result. Nothing covers side effects of a tool call cut off by an interruption (12 of 20). Works with OpenAI Realtime clients after changing the URL and key, with no Inworld Realtime SDK (12 of 15). Total 68.); Rime Prism 73 (Scored as a model API. A 65,536-token context, with `truncation` set to leave out the oldest turns by default, a configurable `retention_ratio`, a `prism_truncation` report on each reply, `prism.conversation.get` to read the stored history and `conversation.item.delete` to trim it (20 of 25). Replies are capped at 120 seconds or 512 tokens and can't be set lower. `max_response_output_tokens`, `temperature` and text-only replies are refused. Up to 16 labels per reply for routing. No lists to page. Partly met, so the earlier 8 is kept (8 of 20). Errors carry `code`, `scope`, `owner`, `param` and a `correlation_id`, and the docs tell clients to branch on code and scope, never on message (20). Client events are matched by `event_id`, a repeated identical tool result is acknowledged, and Prism records late results. Prism doesn't deduplicate requests, a dropped connection doesn't resume, and the client replays history into a new session. Partly met, so the earlier 12 is kept (12 of 20). Short instructions are enough to start, the voice defaults to `astra` and interruption is on. Official SDKs in Python, Node.js, Go and Rust, all alpha, plus LiveKit and Pipecat quickstarts. Partly met while the SDKs are alpha, so the earlier 13 is kept (13 of 15). 20+8+20+12+13 = 73.)