feat(image-gen): add MiniMax image-01 backend plugin - #25451
Closed
ayushere wants to merge 5 commits into
Closed
Conversation
Adds a new image generation backend for MiniMax's image-01 model via https://api.minimax.io/v1/image_generation. - Implements ImageGenProvider with full setup schema, model catalog, and aspect ratio mapping (all 9 MiniMax-native ratios supported) - Downloads generated image from returned URL and caches locally under HERMES_HOME/cache/images/; falls back to raw URL on download failure - Requires MINIMAX_API_KEY; integrates with hermes tools picker - Consistent with existing openai/xai plugin patterns (uses requests, same error_response/success_response conventions) Select via: image_gen.provider: minimax in config.yaml
Contributor
There was a problem hiding this comment.
Pull request overview
Note
Copilot was unable to run its full agentic suite in this review.
Adds a new MiniMax-based image generation backend plugin to the image_gen system.
Changes:
- Introduces
minimaxplugin metadata (plugin.yaml) including required env var. - Implements
MiniMaxImageProviderwith model selection and API calls to MiniMax’simage_generationendpoint. - Adds optional image downloading/caching behavior after receiving a generated image URL.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 6 comments.
| File | Description |
|---|---|
| plugins/image_gen/minimax/plugin.yaml | Declares the new minimax backend plugin and required env configuration. |
| plugins/image_gen/minimax/init.py | Implements the MiniMax image generation provider, config resolution, API request/response handling, and registration. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+131
to
+140
| try: | ||
| resp = requests.post( | ||
| API_ENDPOINT, | ||
| json=payload, | ||
| headers={"Authorization": f"Bearer {api_key}"}, | ||
| timeout=120, | ||
| ) | ||
| resp.raise_for_status() | ||
| data = resp.json() | ||
| except requests.RequestException as exc: |
Comment on lines
+45
to
+46
| # MiniMax natively supports these aspect ratios | ||
| _ASPECT_RATIOS = {"1:1", "16:9", "9:16", "4:3", "3:4", "3:2", "2:3", "21:9"} |
Comment on lines
+120
to
+121
| # Fall back to 1:1 for any ratio not natively supported | ||
| mm_ratio = aspect_ratio if aspect_ratio in _ASPECT_RATIOS else "1:1" |
| image=image_path, | ||
| model=model, | ||
| prompt=prompt, | ||
| aspect_ratio=aspect_ratio, |
Comment on lines
+169
to
+176
| # Download and cache the image locally | ||
| try: | ||
| img_resp = requests.get(image_urls[0], timeout=60) | ||
| img_resp.raise_for_status() | ||
| import base64 | ||
| b64 = base64.b64encode(img_resp.content).decode() | ||
| saved = save_b64_image(b64, prefix="minimax", extension="png") | ||
| image_path = str(saved) |
|
|
||
| import logging | ||
| import os | ||
| from typing import Any, Dict, List, Optional, Tuple |
- Remove unused Optional and Tuple imports; add urllib.parse.urlparse - Catch ValueError (JSONDecodeError) around resp.json() separately from RequestException, returning a clean error_response instead of raising - Return mm_ratio (the ratio actually sent to the API) in all success_response and error_response calls, not the original aspect_ratio which may have been silently remapped to 1:1 - Add SSRF guard: validate image URL is HTTPS and host matches an explicit MiniMax-controlled allowlist before issuing requests.get; fall back to returning the raw URL and log a warning on mismatch
alt-glitch
added
type/feature
Collaborator
…F allowlist
- subject_reference: character-consistent generation from a portrait URL
or base64 data URL. Passes subject_reference=[{type:character,image_file}]
to the MiniMax image_generation API when provided via kwargs.
- prompt_optimizer: forward prompt_optimizer=True to API when caller requests it.
- SSRF allowlist: add two additional MiniMax image CDN hosts seen in production
(aliyuncs accelerate + minimax-algeng) so generated images cache locally.
- Expose subject_reference in image_generate tool schema so the agent can use
character-consistent generation without manual API wiring.
- Pass **extra_kwargs through _dispatch_to_plugin_provider so future providers
can receive additional parameters.
13 tasks
This was referenced Jul 12, 2026
Contributor
|
Thanks for the MiniMax image-generation implementation and for following the existing This automated hermes-sweeper review is closing the PR under the standing third-party provider-integration policy: vendor-specific integrations must ship as standalone plugins, not as new in-tree directories under
Please publish or continue this work as a standalone MiniMax image-generation plugin repository, then share it in Automated hermes-sweeper review; closed as not planned under standing policy Closed as not-planned per standing maintainer policy ( |
teknium1
added
the
sweeper:not-planned
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a new image generation backend plugin for MiniMax's
image-01model.MiniMax is already a fully-supported LLM provider in Hermes, but had no image generation support. This fills the gap using the same plugin architecture as the existing
openai,openai-codex, andxaibackends.What this adds
plugins/image_gen/minimax/__init__.py— fullImageGenProviderimplementationPOST https://api.minimax.io/v1/image_generation1:1,16:9,9:16,4:3,3:4,3:2,2:3,21:9HERMES_HOME/cache/images/; falls back to raw URL on download failureget_setup_schema()forhermes toolspicker integrationimage_gen.minimax.model(orMINIMAX_IMAGE_MODELenv var)plugins/image_gen/minimax/plugin.yaml— manifest following existing plugin conventionsUsage
Requires
MINIMAX_API_KEYin~/.hermes/.env.Notes
plugins/image_gen/xai/andplugins/image_gen/openai/requests(consistent with other plugins, noturllib)1002rate limit,1004auth,1008balance,1026content policy)