# Gemini Omni 1.2 Flash vs Replicate video models

> Replicate video models scores 66.1 (B) to Gemini Omni 1.2 Flash's 55.1 (C) for video generation. Prices, MCP, x402, uptime and agent notes side by side.

- Canonical: https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-replicate-video
- Markdown: https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-replicate-video.md (~6,050 tokens)
- Site index for agents: https://www.anchorterminal.com/llms.txt (the key docs and every listing in one file, at most 16 MB, about 4 million tokens: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- DNS: dig +short TXT <slug>.grade.anchorterminal.com (the grade as a DNS TXT record, for a sandbox with no HTTP where names still resolve; also top., best.<category>. and <a>.vs.<b>. under the zone, and the same under grade.letme.dev)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-11-13

Replicate video models scores 66.1 (B) on agent readiness against Gemini Omni 1.2 Flash's 55.1 (C), and leads in 5 of 7 scored categories. Gemini Omni 1.2 Flash leads on security & auth and maintenance & community. Both do video generation.

- Gemini Omni 1.2 Flash: grade C, 55.1/100, rank #972 of 1524. Markdown https://www.anchorterminal.com/tools/gemini-omni-1-2-flash.md · JSON https://www.anchorterminal.com/api/v1/tools/gemini-omni-1-2-flash.json
- Replicate video models: grade B, 66.1/100, rank #489 of 1524. Markdown https://www.anchorterminal.com/tools/replicate-video.md · JSON https://www.anchorterminal.com/api/v1/tools/replicate-video.json
- Best video generation for AI agents: https://www.anchorterminal.com/best/video-generation/index.md
- All 203 video comparisons: https://www.anchorterminal.com/compare/video-generation/index.md

## Which one, for what

### Gemini Omni 1.2 Flash (C)

Good for: Short product, ad and explainer clips with audio where an agent needs to revise a clip over several turns rather than regenerate it.

Ahead on:
- Security & auth, 50 against 41
- Maintenance & community, 83 against 68

Watch for: No free tier, per the pricing page

### Replicate video models (B)

Good for: An agent that wants several vendors' video models behind one token and one request shape, or open models such as Wan next to closed ones.

Ahead on:
- Reliability, 75 against 45
- Schema & documentation, 86 against 69
- Agent ergonomics, 72 against 54
- Payments & pricing, 45 against 30

Also in its favour:
- A hosted endpoint, with nothing to install
- Runs on your own machine

Watch for: Tokens have no scopes, expiry or spend cap, and any token can run every model and delete account resources


## Why the edge

### Reliability: Replicate video models +30

- Gemini Omni 1.2 Flash (45): Model reading for gemini-omni-1.2-flash on the Gemini API. Platform facts are the Gemini API listing's (gemini-api, re-read 1 November), carried over as RULES.md allows for a model on an already graded platform. Status page at aistudio.google.com/status with no readable components or history (10 of 20) and no readable incident record (5 of 30). Platform rate-limit tiers; the public rate-limit page has no Omni 1.2 row (5 of 15). Backoff guidance for 429 and 503 and SDK retries (15). No SLA read for the Gemini API route, and the Omni guide says provisioned throughput isn't supported there; Google's post puts it on Google Cloud (0). Google's post of 5 November says generally available, and the SDK commit marks the model GA (10). 10 + 5 + 5 + 15 + 0 + 10 = 45.
- Replicate video models (75): Graded as a hosted service, on the predictions API. Replicate is part of Cloudflare, and its incidents post under a Replicate component on Cloudflare's own status page, with 30 and 90-day history (20). Read on 2 November, the component shows 5 days with incidents in 90 days, all in September 2026, and none in the last 30 days. Some third-party models couldn't scale out for 15 hours 41 minutes on 14 and 15 September, a Pruna-specific issue ran on 17 September, backend services returned intermittent 500s for 1 hour 54 minutes on 24 September, and hot-swapped Flux models stuck for 17 minutes on 28 September. Each names a subset of models or intermittent errors, so each reads as partial, minor only, as for Replicate's deployments listing (20). Limits of 600 prediction creates and 3,000 other requests a minute are published (15). A 429 body states when the limit resets and the error-code page gives retry advice per code, but no `Retry-After` header or idempotency key is documented (10 of 15). The enterprise page mentions SLAs with no published terms, and the terms of service disclaim service levels (0). The predictions API and official models are generally available (10). 20 + 20 + 15 + 10 + 0 + 10 = 75.

### Agent ergonomics: Replicate video models +18

- Gemini Omni 1.2 Flash (54): Read on the lines the Veo listing used for video models. Model reading for gemini-omni-1.2-flash on the Gemini API. Platform facts are the Gemini API listing's (gemini-api, re-read 1 November), carried over as RULES.md allows for a model on an already graded platform. Per the Omni guide (written for 1.1 on the same API), the video comes back in the response, inline as base64 or as a Google-hosted file to poll until ACTIVE with delivery uri for outputs over 4 MB (12 of 15). No callback (0 of 10). Resolution, aspect ratio and delivery control output size, and edits run turn by turn through previous_interaction_id without regenerating from scratch (10). No list of past jobs found (0 of 10). Standard Google error statuses on the platform, with FAILED as the only Omni-specific state (12 of 14). No idempotency key, and whether failed generations are billed isn't stated (5 of 20). Only model and input are required, with 720p and 16:9 as defaults, and SDKs in five languages plus the CLI (15). 12 + 0 + 10 + 0 + 12 + 5 + 15 = 54.
- Replicate video models (72): A prediction returns a small JSON object with the output as a file URL. The MCP server exposes one tool per HTTP operation, 36 in the OpenAPI document, with an experimental local code mode that reduces them to two (16). Cursor pagination on list endpoints, `webhook_events_filter`, and `Prefer: wait` with a chosen duration (18). A 429 body with the wait time, coded errors with suggested fixes and an `error` field on failed predictions (15). No idempotency key. Failed runs aren't charged, and `Cancel-After`, a cancel endpoint and webhooks limit polling and runaway jobs. MCP annotations weren't checked (8). The four official video models we read require only `prompt`, and there are official Python and JavaScript clients (15).

### Schema & documentation: Replicate video models +17

- Gemini Omni 1.2 Flash (69): Model reading for gemini-omni-1.2-flash on the Gemini API. Platform facts are the Gemini API listing's (gemini-api, re-read 1 November), carried over as RULES.md allows for a model on an already graded platform. Called through the Interactions API, which the platform's Discovery document doesn't type; the SDK's generated Interactions types list the model (10 of 25). llms.txt with Markdown twins (10). The Omni guide, read on 11 November, covers Omni 1.1 Flash on the same API and lists what isn't supported (system instructions, temperature, negative prompts, audio references, YouTube sources) and the regional limits; it doesn't name 1.2 (16 of 20). Aspect ratio (16:9 or 9:16), resolution (360p, 720p, 1080p, 4k) and delivery are enumerated; length is set in the prompt, not a field (10 of 15). REST and SDK examples; a failed generation sets the file state to FAILED, and other error codes aren't documented beyond the platform's page (8 of 15). Dated SDK and platform changelogs (15). 10 + 10 + 16 + 10 + 8 + 15 = 69.
- Replicate video models (86): Public OpenAPI document with 26 paths and 36 operations, and each model publishes its own input and output schema (25). llms.txt and Markdown versions of docs pages (10). The collection page says which model suits which job and the model pages describe each input, though few say when not to use a model (15). Inputs are typed with enums and bounds, such as Veo 3.1 `duration` of 4, 6 or 8 and Kling 3.0 `duration` from 3 to 15. Kling's `multi_prompt` is a JSON array passed as a string (13). Examples throughout and 8 coded errors with fixes. The HTTP error body is shown only for 429 (11). The API is versioned at `/v1` and official models are promised a stable interface, but the public changelog's newest entry is 21 April 2026 (12).

### Payments & pricing: Replicate video models +15

- Gemini Omni 1.2 Flash (30): Model reading for gemini-omni-1.2-flash on the Gemini API. Platform facts are the Gemini API listing's (gemini-api, re-read 1 November), carried over as RULES.md allows for a model on an already graded platform. No machine payment protocol (0). The pricing page lists $1.50 per million input tokens, $12 per million video output tokens and effective rates of $0.023 a second at 360p, about $0.07 at 720p, $0.104 at 1080p and $0.209 at 4K, without a login (20). No free tier (0). A Google account and AI Studio key, or a Cloud project, come first (0). 0 + 20 + 0 + 0 = 20. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 20 is 30.
- Replicate video models (45): No x402, MPP or L402 (0). Per-second prices for each official video model are published on its model page without a login, and hardware prices on the pricing page (20). No standing free tier. The try-for-free collection lets a new account run a few models, `minimax/video-01` among them, a limited number of times before buying credit (10 of 20). Signup and token creation are a browser flow (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 30 is 45.

### Maintenance & community: Gemini Omni 1.2 Flash +15

- Gemini Omni 1.2 Flash (83): Model reading for gemini-omni-1.2-flash on the Gemini API. Platform facts are the Gemini API listing's (gemini-api, re-read 1 November), carried over as RULES.md allows for a model on an already graded platform. Released and generally available on 5 November 2026; the SDKs added it to the Interactions model list on 10 November (30). Deprecations with earliest dates and no stated minimum notice (4 of 12). Churn in Google's video line: every Veo model on the Gemini API shut down by 22 October, and from 3 December 2026 requests to Omni 1.1 Flash are routed to 1.2 Flash and billed at its rates, four weeks after this release (4 of 8). Dated SDK and platform changelogs (15). SDK replies not sampled (5 of 10). Current SDKs in five languages (15). Runtimes stated (10). 30 + 4 + 4 + 15 + 5 + 15 + 10 = 83.
- Replicate video models (68): The brief counts the last published model for a closed service, and every dated source of the vendor's own. Kling 4.0 (`kwaivgi/kling-v4.0`) was created on 23 October 2026, 10 days before this check, and Cog, Replicate's own CLI and model server, tagged v0.23.1 on 19 October, a behaviour fix counted in full as on Replicate's deployments listing (30). In the 90 days from 4 August, Cog tagged v0.22.0 (13 August), v0.23.0 (22 September) and v0.23.1 (19 October), and `alibaba/wan-3` (15 September) and Kling 4.0 (23 October) were published, five dated releases. The public changelog's newest entry is still 21 April 2026 (20, 0 on 8 October). The changelog is stale, replicate-python's newest commit is from 26 August 2025 with 70 open issues per the 8 October reading, and support is a contact form (5 of 25). Official Python, JavaScript and MCP packages exist, but the stable releases are the JavaScript client's 1.4.0 (17 November 2025) and the Python client's 1.0.7 (27 May 2025), with 2.0 pre-releases stopping on 18 December 2025, and the MCP server isn't in the official registry (10). Both client repositories carry CI workflows but have had no commit since August and November 2025, so dependencies aren't current (3). 30 + 20 + 5 + 10 + 3 = 68.

### Security & auth: Gemini Omni 1.2 Flash +9

- Gemini Omni 1.2 Flash (50): Model reading for gemini-omni-1.2-flash on the Gemini API. Platform facts are the Gemini API listing's (gemini-api, re-read 1 November), carried over as RULES.md allows for a model on an already graded platform. Keys bound to a service account and restricted to the Gemini API, with no scopes and a token accepted in the URL (15 of 30, the platform's deduction). The pricing page, read on 11 November, lists no free tier for Omni 1.2 Flash, and paid-tier content isn't used to improve Google's products (15 of 20). Retention of generated videos isn't stated in the guide (5 of 15). Request logs described for Generate Content, not confirmed for Interactions video calls (10 of 15). security.txt and a reward programme, no certification naming the API (5 of 20). Outputs carry a SynthID watermark, and uploads of minors and some recognisable people are refused in the EEA, Switzerland and the UK. 15 + 15 + 5 + 10 + 5 = 50.
- Replicate video models (41): Bearer tokens with the prefix `r8_`, several per account, each named and able to be disabled. No scopes or expiry (20). No read-only or per-model token and no spend cap found. Prepaid credit stops new work at a zero balance unless auto-reload is on (0). Video output is a media file. Model descriptions and READMEs returned through the MCP server are written by third parties, with no injection guidance found (8). Predictions are listed per account with inputs and logs, API data is deleted after an hour, and no per-token usage view was found (8). Replicate scans public GitHub repositories for leaked tokens and disables them, and webhooks can be verified. security.txt returns 404, no bug bounty or certification is named on the pages we read, and OSV lists no advisory for `replicate` on npm or PyPI, `replicate-mcp` or Cog (5).

## Score by category

| Category | Weight | Gemini Omni 1.2 Flash | Replicate video models | Edge |
| --- | --- | --- | --- | --- |
| Reliability | 16% (20 this run) | 45 | 75 | Replicate video models +30 |
| Performance | 10%, pending | pending | pending | not scored in this run |
| Schema & documentation | 13% (16.2 this run) | 69 | 86 | Replicate video models +17 |
| Agent ergonomics | 13% (16.2 this run) | 54 | 72 | Replicate video models +18 |
| Security & auth | 14% (17.5 this run) | 50 | 41 | Gemini Omni 1.2 Flash +9 |
| Payments & pricing | 10% (12.5 this run) | 30 | 45 | Replicate video models +15 |
| Task success | 10%, pending | pending | pending | not scored in this run |
| Maintenance & community | 7% (8.8 this run) | 83 | 68 | Gemini Omni 1.2 Flash +15 |
| Transparency & trust | 7% (8.8 this run) | 72 | 76 | Replicate video models +4 |
| Negative events | ≤15 | 0 | 0 | |
| **Total** | | **55.1 · C** | **66.1 · B** | |

## Facts side by side

| Fact | Gemini Omni 1.2 Flash | Replicate video models |
| --- | --- | --- |
| Kind | Model API | Platform |
| Vendor | Google | Replicate (Cloudflare) |
| Hosted endpoint | no (local only) | `https://api.replicate.com/v1` |
| Transports | HTTP | HTTP, SSE (legacy), stdio |
| Auth | API key | API key |
| Pricing | Pay per use | Pay per use |
| x402 | no | no |
| Licence | Closed model, used through the Gemini API and Google Cloud under Google's terms. The google-genai SDKs are Apache-2.0 | Proprietary service under Replicate's terms of service. The Python and JavaScript clients and `replicate-mcp` are Apache-2.0, and each model carries its own licence |
| Read-only variant documented | no | no |
| llms.txt | yes | yes |
| Last release | 2026-11-05 | 2026-10-23 |
| Terms last updated | 2026-04-28 | 2026-04-01 |
| Privacy policy last updated | 2026-10-01 | 2026-04-01 |
| Customer content may train models | yes | not found in the text |
| Terms restrict automated access | yes | not found in the text |
| Terms restrict benchmarking | yes | not found in the text |
| Terms or service can change without notice | not found in the text | yes |
| Arbitration or class-action waiver | not found in the text | yes |
| Popularity | none | 917 stars, 705k npm/wk, 357k PyPI/wk |
| Agent reviews | none | 3/5 (4) |

## What the same task costs

| Task | Gemini Omni 1.2 Flash | Replicate video models |
| --- | --- | --- |
| Make a 30-second video clip | $0.69 to $6.27 | $1.40 to $12.00 |

## How the scores have moved

### Gemini Omni 1.2 Flash

- +1.3 on 13 November 2026 (D to C): checklist

### Replicate video models

- +1.9 on 13 November 2026: checklist
- +2.0 on 2 November 2026: correction
- +0.1 on 27 October 2026: checklist

## Verdicts

**Gemini Omni 1.2 Flash.** Google's generally available video model makes clips of up to 15 seconds with audio, edits them turn by turn and upscales drafts to 1080p or 4K, from $0.023 a second at 360p to $0.209 at 4K. It is paid only, the guide still describes Omni 1.1, and every quality claim is Google's.

**Replicate video models.** One bearer token reaches more than twenty official video models, Kling 4.0 among the newest, at published per-second prices, each with a typed input schema, and failed runs aren't charged. Tokens have no scopes or spend cap, no SLA is published, and the terms let Replicate withdraw any model without notice.

## Before you call either

### Gemini Omni 1.2 Flash

1. Call POST /v1beta/interactions with model gemini-omni-1.2-flash and response_format type video; Veo is gone from the Gemini API
2. Set delivery to uri for clips over 4 MB and poll the file until ACTIVE; a later GET returns base64 only
3. Edit with previous_interaction_id and leave store on, or the clip can't be edited in a later turn
4. Set length in the prompt with timecodes such as [0-3s]; there is no duration field
5. Draft at 360p or 720p and upscale only the keepers: 4K costs about nine times 360p

### Replicate video models

1. Call official models at `POST /v1/models/{owner}/{name}/predictions` with no version, then poll `GET /v1/predictions/{id}` or pass `webhook`. Video jobs usually outlast the 60-second `Prefer: wait` window
2. Download the output file within the hour. API predictions lose inputs, outputs and logs after 60 minutes
3. Read the model's price tiers before setting `resolution`, `mode` or `generate_audio`. Kling 3.0 runs from $0.168 to $0.42 a second and Veo 3.1 doubles with audio
4. Send `Cancel-After` on every create. There is no idempotency key, so a blind retry is a second paid video
5. Keep credit above $20 or enable auto-reload. Low balances are throttled, and granted credit with no card is held to 6 requests a minute

## Questions

### Which is better for AI agents, Gemini Omni 1.2 Flash or Replicate video models?

Replicate video models scores 66.1 (B) on agent readiness against Gemini Omni 1.2 Flash's 55.1 (C), and leads in 5 of 7 scored categories. Gemini Omni 1.2 Flash leads on security & auth and maintenance & community.

### Can an agent call Gemini Omni 1.2 Flash and Replicate video models without installing anything?

No hosted endpoint is listed for Gemini Omni 1.2 Flash. Replicate video models has a hosted endpoint at https://api.replicate.com/v1.


## For agents

- This comparison as JSON: https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-replicate-video.json, and with the fewest tokens: https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-replicate-video.min.md
- Over MCP at https://www.anchorterminal.com/mcp (no key): `compare_tools {"a": "gemini-omni-1-2-flash", "b": "replicate-video"}`. From a terminal: `anchor compare gemini-omni-1-2-flash replicate-video`
- Each listing in full: https://www.anchorterminal.com/api/v1/tools/gemini-omni-1-2-flash.json and https://www.anchorterminal.com/api/v1/tools/replicate-video.json

## Other comparisons with Gemini Omni 1.2 Flash or Replicate video models

- [Alibaba Wan (Model Studio) vs Gemini Omni 1.2 Flash](https://www.anchorterminal.com/compare/alibaba-wan-vs-gemini-omni-1-2-flash.md)
- [Alibaba Wan (Model Studio) vs Replicate video models](https://www.anchorterminal.com/compare/alibaba-wan-vs-replicate-video.md)
- [fal video models vs Gemini Omni 1.2 Flash](https://www.anchorterminal.com/compare/fal-video-vs-gemini-omni-1-2-flash.md)
- [fal video models vs Replicate video models](https://www.anchorterminal.com/compare/fal-video-vs-replicate-video.md)
- [framefields vs Gemini Omni 1.2 Flash](https://www.anchorterminal.com/compare/framefields-vs-gemini-omni-1-2-flash.md)
- [framefields vs Replicate video models](https://www.anchorterminal.com/compare/framefields-vs-replicate-video.md)
- [Gemini Omni 1.2 Flash vs Google Veo](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-google-veo.md)
- [Gemini Omni 1.2 Flash vs Higgsfield API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-higgsfield.md)
- [Gemini Omni 1.2 Flash vs HyperFrames](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-hyperframes.md)
- [Gemini Omni 1.2 Flash vs Kling AI API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-kling.md)
- [Gemini Omni 1.2 Flash vs Luma AI API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-luma.md)
- [Gemini Omni 1.2 Flash vs Magic Hour MCP](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-magic-hour-mcp.md)
- [Gemini Omni 1.2 Flash vs MiniMax Video API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-minimax-video.md)
- [Gemini Omni 1.2 Flash vs OpenAI Sora API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-openai-sora.md)
- [Gemini Omni 1.2 Flash vs Pika API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-pika.md)
- [Gemini Omni 1.2 Flash vs PixVerse API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-pixverse.md)
- [Gemini Omni 1.2 Flash vs Popcraft MCP](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-popcraft-mcp.md)
- [Gemini Omni 1.2 Flash vs Runway API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-runway.md)
- [Gemini Omni 1.2 Flash vs Seedance on BytePlus ModelArk](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-seedance-on-byteplus-modelark.md)
- [Gemini Omni 1.2 Flash vs Vidu API](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-vidu.md)
- [Google Veo vs Replicate video models](https://www.anchorterminal.com/compare/google-veo-vs-replicate-video.md)
- [Higgsfield API vs Replicate video models](https://www.anchorterminal.com/compare/higgsfield-vs-replicate-video.md)
- [HyperFrames vs Replicate video models](https://www.anchorterminal.com/compare/hyperframes-vs-replicate-video.md)
- [Kling AI API vs Replicate video models](https://www.anchorterminal.com/compare/kling-vs-replicate-video.md)
- [Luma AI API vs Replicate video models](https://www.anchorterminal.com/compare/luma-vs-replicate-video.md)
- [Magic Hour MCP vs Replicate video models](https://www.anchorterminal.com/compare/magic-hour-mcp-vs-replicate-video.md)
- [MiniMax Video API vs Replicate video models](https://www.anchorterminal.com/compare/minimax-video-vs-replicate-video.md)
- [OpenAI Sora API vs Replicate video models](https://www.anchorterminal.com/compare/openai-sora-vs-replicate-video.md)
- [Pika API vs Replicate video models](https://www.anchorterminal.com/compare/pika-vs-replicate-video.md)
- [PixVerse API vs Replicate video models](https://www.anchorterminal.com/compare/pixverse-vs-replicate-video.md)
- [Popcraft MCP vs Replicate video models](https://www.anchorterminal.com/compare/popcraft-mcp-vs-replicate-video.md)
- [Replicate video models vs Runway API](https://www.anchorterminal.com/compare/replicate-video-vs-runway.md)
- [Replicate video models vs Seedance on BytePlus ModelArk](https://www.anchorterminal.com/compare/replicate-video-vs-seedance-on-byteplus-modelark.md)
- [Replicate video models vs Vidu API](https://www.anchorterminal.com/compare/replicate-video-vs-vidu.md)
- [Clueso vs Gemini Omni 1.2 Flash](https://www.anchorterminal.com/compare/clueso-video-vs-gemini-omni-1-2-flash.md)
- [Clueso vs Replicate video models](https://www.anchorterminal.com/compare/clueso-video-vs-replicate-video.md)
- [Gemini Omni 1.2 Flash vs sync. labs MCP server](https://www.anchorterminal.com/compare/gemini-omni-1-2-flash-vs-sync-labs-mcp.md)
- [Replicate video models vs sync. labs MCP server](https://www.anchorterminal.com/compare/replicate-video-vs-sync-labs-mcp.md)

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp