feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large) by teknium1 · Pull Request #33236 · NousResearch/hermes-agent

teknium1 · 2026-05-27T12:34:47Z

Summary

Adds Krea as a built-in image_gen provider — Krea 2 Medium ($0.03) and Krea 2 Large ($0.06). Auto-discovered like the other image_gen plugins; users select it via hermes tools → Image Generation → Krea.

Changes

plugins/image_gen/krea/ — new kind: backend plugin (plugin.yaml + __init__.py)
hermes_cli/config.py — KREA_API_KEY added to OPTIONAL_ENV_VARS so hermes setup prompts for it
tests/plugins/image_gen/test_krea_provider.py — 31 tests

Design notes

Krea's API is asynchronous: submit returns a job_id, then poll GET /jobs/{id} until terminal. The provider hides this behind the synchronous ImageGenProvider.generate() contract — submit, poll every 2s with light backoff (max 5s), 3-minute ceiling matching Krea's hosted-tool timeout.

Result URLs are materialised to $HERMES_HOME/cache/images/ so the gateway doesn't 404 on expired Krea CDN links — same defense the xAI provider uses (#26942).

Aspect ratio map: landscape→16:9, square→1:1, portrait→9:16. Resolution: 1K (Krea's only current option).

Kwarg passthrough: seed, creativity (raw/low/medium/high), styles, image_style_references (capped at 10), moodboards (capped at 1) — matches Krea's per-request limits. Unknown kwargs are ignored.

Config

image_gen:
  provider: krea
  krea:
    model: krea-2-medium    # or krea-2-large
    creativity: medium      # raw | low | medium | high

Env overrides: KREA_API_KEY (required), KREA_IMAGE_MODEL.

Validation

Check	Result
`tests/plugins/image_gen/test_krea_provider.py`	31/31
Full image_gen suite + picker + tools_config	211/211
E2E plugin registration via `_ensure_plugins_discovered()`	passes
E2E picker row injection in `_plugin_image_gen_providers()`	passes
E2E active-provider selection (`image_gen.provider=krea`)	passes
E2E model routing (`krea-2-large` → `/krea-2/large` endpoint)	passes

Not live-tested yet — pending Krea API key issuance after workspace registration.

Refs: https://docs.krea.ai/developers/krea-2/overview, https://docs.krea.ai/api-reference/krea/krea-2-large

Infographic

New built-in image_gen backend wrapping Krea's Krea 2 foundation image model family. Auto-discovered like the other image_gen plugins and appears in 'hermes tools' → Image Generation → Krea. Krea's API is asynchronous — submit returns a job_id, poll /jobs/{id} until terminal. The provider hides that behind the synchronous ImageGenProvider.generate() contract: submit, poll every 2s with light backoff (max 5s), 3-minute ceiling matching Krea's hosted-tool timeout. Result URL is materialised to $HERMES_HOME/cache/images/ to avoid CDN-expiry 404s downstream (same fix as xAI #26942). Models: - krea-2-medium (default — Krea's 'start here' recommendation) - krea-2-large Aspect ratios map landscape→16:9, square→1:1, portrait→9:16. Resolution: 1K (Krea's only current option). Kwarg passthrough: seed, creativity (raw/low/medium/high), styles, image_style_references (capped 10), moodboards (capped 1) — matches Krea's per-request limits. Unknown kwargs are ignored. Config knobs (config.yaml): image_gen.provider: krea image_gen.krea.model: krea-2-medium | krea-2-large image_gen.krea.creativity: raw | low | medium | high Env overrides: KREA_API_KEY (required), KREA_IMAGE_MODEL. KREA_API_KEY is registered in OPTIONAL_ENV_VARS so 'hermes setup' prompts for it. 31 new tests; image_gen suite + picker + tools_config: 211/211.

github-actions · 2026-05-27T12:35:30Z

🔎 Lint report: `hermes/hermes-187cc3e8` vs `origin/main`

ruff

Total: 0 on HEAD, 0 on base (➖ 0)

🆕 New issues: none

✅ Fixed issues: none

Unchanged: 0 pre-existing issues carried over.

ty (type checker)

Total: 9491 on HEAD, 9488 on base (🆕 +3)

🆕 New issues (2):

Rule	Count
`invalid-assignment`	1
`unresolved-import`	1

First entries

tests/plugins/image_gen/test_krea_provider.py:514: [invalid-assignment] invalid-assignment: Object of type `MagicMock` is not assignable to attribute `raise_for_status` of type `def raise_for_status(self) -> Unknown`
tests/plugins/image_gen/test_krea_provider.py:9: [unresolved-import] unresolved-import: Cannot resolve imported module `pytest`

✅ Fixed issues: none

Unchanged: 5001 pre-existing issues carried over.

Diagnostics are surfaced as warnings — this check never fails the build.

albertsalgueda · 2026-05-27T14:28:27Z

+    "KREA_API_KEY": {
+        "description": "Krea API key for Krea 2 image generation (Medium + Large)",
+        "prompt": "Krea API key",
+        "url": "https://www.krea.ai/api-keys",


This URL currently returns 404. Krea’s API docs direct users to https://www.krea.ai/app/api/tokens for API token creation, so setup should point there instead.

Good catch — fixed in 98cd760. Switched to https://www.krea.ai/app/api/tokens (verified 303 → tokens page; old URL was 404).

albertsalgueda · 2026-05-27T14:29:00Z

+            try:
+                poll_resp = requests.get(job_url, headers=poll_headers, timeout=30)
+                poll_resp.raise_for_status()
+            except requests.HTTPError as exc:


This retries every polling HTTP error until the 180s deadline, including permanent failures like 401, 402, 403, or 404. The comment says transient 5xx, but the code does not distinguish status classes, so an auth/billing/not-found issue can make image_generate hang for three minutes. Please fail fast for non-retryable 4xx responses and only retry transient statuses such as 408/409/425/429/5xx if those are expected from Krea.

Fixed in 98cd760. Added _RETRYABLE_POLL_STATUSES = {408, 409, 425, 429, 500, 502, 503, 504} and changed the poll loop to fail fast on anything outside that set — so 401/402/403/404 surface immediately instead of burning the 180s deadline. Added 5 tests covering fail-fast on 401/403/404 and retry on 429/503.

- Update KREA_API_KEY setup URL to the canonical token-creation page (https://www.krea.ai/app/api/tokens). The previous URL returned 404. - Fail fast on non-retryable HTTP statuses during poll. The previous loop retried every HTTPError for the full 180s deadline, so an auth (401), billing (402), forbidden (403), or not-found (404) response would make image_generate hang for three minutes. Only retry transient statuses (408/409/425/429/5xx); surface everything else immediately. - Add 5 tests covering fail-fast on 401/403/404 and retry on 429/503.

Three call sites linked users to dashboard pages that don't exist: - hermes_cli/config.py: https://www.krea.ai/app/api/tokens - plugins/image_gen/krea/__init__.py get_setup_schema: https://www.krea.ai/api-keys - plugins/image_gen/krea/__init__.py auth_required error: https://www.krea.ai/api-keys Per Krea's own docs (https://docs.krea.ai/developers/api-keys-and-billing), the real dashboard URL is https://www.krea.ai/settings/api-tokens. All three sites now point there.

fal announced Krea 2 day-0 as an official API partner on 2026-05-27. Add both variants to the FAL_MODELS catalog so they appear in the 'hermes tools' model picker alongside flux-2, gpt-image, nano-banana, etc. Users who already bill through FAL or Nous Portal subscription can now use Krea without registering directly with Krea. Model IDs (as listed in fal's launch announcement): fal-ai/krea/v2/medium/text-to-image — $0.030 / image fal-ai/krea/v2/large/text-to-image — $0.060 / image Both share the same parameter schema: - aspect_ratio (1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 2:3, 9:16) mapped from our 3 abstract ratios via size_style='aspect_ratio' - creativity (raw|low|medium|high; default medium) - seed (reproducibility) - image_style_references (up to 10 per Krea's API spec) No num_inference_steps / guidance_scale / num_images — Krea 2 does not expose those, and the supports-set filter strips them defensively if the agent ever passes them. This is the FAL-routed variant. The separate native-Krea-API plugin shipped in PR #33236 (plugins/image_gen/krea/) remains available for users who want to bill directly through Krea's API with their own key. Both routes converge on the same underlying model. Nous Portal managed-FAL gateway: this commit makes the model IDs known to the catalog and the picker. The Portal team will need to allowlist these two endpoint slugs on the fal-queue origin server-side for them to flow through the managed billing path.

…usResearch#33236) * feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large) New built-in image_gen backend wrapping Krea's Krea 2 foundation image model family. Auto-discovered like the other image_gen plugins and appears in 'hermes tools' → Image Generation → Krea. Krea's API is asynchronous — submit returns a job_id, poll /jobs/{id} until terminal. The provider hides that behind the synchronous ImageGenProvider.generate() contract: submit, poll every 2s with light backoff (max 5s), 3-minute ceiling matching Krea's hosted-tool timeout. Result URL is materialised to $HERMES_HOME/cache/images/ to avoid CDN-expiry 404s downstream (same fix as xAI NousResearch#26942). Models: - krea-2-medium (default — Krea's 'start here' recommendation) - krea-2-large Aspect ratios map landscape→16:9, square→1:1, portrait→9:16. Resolution: 1K (Krea's only current option). Kwarg passthrough: seed, creativity (raw/low/medium/high), styles, image_style_references (capped 10), moodboards (capped 1) — matches Krea's per-request limits. Unknown kwargs are ignored. Config knobs (config.yaml): image_gen.provider: krea image_gen.krea.model: krea-2-medium | krea-2-large image_gen.krea.creativity: raw | low | medium | high Env overrides: KREA_API_KEY (required), KREA_IMAGE_MODEL. KREA_API_KEY is registered in OPTIONAL_ENV_VARS so 'hermes setup' prompts for it. 31 new tests; image_gen suite + picker + tools_config: 211/211. * fix(image_gen/krea): address review feedback - Update KREA_API_KEY setup URL to the canonical token-creation page (https://www.krea.ai/app/api/tokens). The previous URL returned 404. - Fail fast on non-retryable HTTP statuses during poll. The previous loop retried every HTTPError for the full 180s deadline, so an auth (401), billing (402), forbidden (403), or not-found (404) response would make image_generate hang for three minutes. Only retry transient statuses (408/409/425/429/5xx); surface everything else immediately. - Add 5 tests covering fail-fast on 401/403/404 and retry on 429/503. * fix(krea): point users at the real API token dashboard URL Three call sites linked users to dashboard pages that don't exist: - hermes_cli/config.py: https://www.krea.ai/app/api/tokens - plugins/image_gen/krea/__init__.py get_setup_schema: https://www.krea.ai/api-keys - plugins/image_gen/krea/__init__.py auth_required error: https://www.krea.ai/api-keys Per Krea's own docs (https://docs.krea.ai/developers/api-keys-and-billing), the real dashboard URL is https://www.krea.ai/settings/api-tokens. All three sites now point there.

…search#33506) fal announced Krea 2 day-0 as an official API partner on 2026-05-27. Add both variants to the FAL_MODELS catalog so they appear in the 'hermes tools' model picker alongside flux-2, gpt-image, nano-banana, etc. Users who already bill through FAL or Nous Portal subscription can now use Krea without registering directly with Krea. Model IDs (as listed in fal's launch announcement): fal-ai/krea/v2/medium/text-to-image — $0.030 / image fal-ai/krea/v2/large/text-to-image — $0.060 / image Both share the same parameter schema: - aspect_ratio (1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 2:3, 9:16) mapped from our 3 abstract ratios via size_style='aspect_ratio' - creativity (raw|low|medium|high; default medium) - seed (reproducibility) - image_style_references (up to 10 per Krea's API spec) No num_inference_steps / guidance_scale / num_images — Krea 2 does not expose those, and the supports-set filter strips them defensively if the agent ever passes them. This is the FAL-routed variant. The separate native-Krea-API plugin shipped in PR NousResearch#33236 (plugins/image_gen/krea/) remains available for users who want to bill directly through Krea's API with their own key. Both routes converge on the same underlying model. Nous Portal managed-FAL gateway: this commit makes the model IDs known to the catalog and the picker. The Portal team will need to allowlist these two endpoint slugs on the fal-queue origin server-side for them to flow through the managed billing path.

…usResearch#33236) * feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large) New built-in image_gen backend wrapping Krea's Krea 2 foundation image model family. Auto-discovered like the other image_gen plugins and appears in 'hermes tools' → Image Generation → Krea. Krea's API is asynchronous — submit returns a job_id, poll /jobs/{id} until terminal. The provider hides that behind the synchronous ImageGenProvider.generate() contract: submit, poll every 2s with light backoff (max 5s), 3-minute ceiling matching Krea's hosted-tool timeout. Result URL is materialised to $HERMES_HOME/cache/images/ to avoid CDN-expiry 404s downstream (same fix as xAI NousResearch#26942). Models: - krea-2-medium (default — Krea's 'start here' recommendation) - krea-2-large Aspect ratios map landscape→16:9, square→1:1, portrait→9:16. Resolution: 1K (Krea's only current option). Kwarg passthrough: seed, creativity (raw/low/medium/high), styles, image_style_references (capped 10), moodboards (capped 1) — matches Krea's per-request limits. Unknown kwargs are ignored. Config knobs (config.yaml): image_gen.provider: krea image_gen.krea.model: krea-2-medium | krea-2-large image_gen.krea.creativity: raw | low | medium | high Env overrides: KREA_API_KEY (required), KREA_IMAGE_MODEL. KREA_API_KEY is registered in OPTIONAL_ENV_VARS so 'hermes setup' prompts for it. 31 new tests; image_gen suite + picker + tools_config: 211/211. * fix(image_gen/krea): address review feedback - Update KREA_API_KEY setup URL to the canonical token-creation page (https://www.krea.ai/app/api/tokens). The previous URL returned 404. - Fail fast on non-retryable HTTP statuses during poll. The previous loop retried every HTTPError for the full 180s deadline, so an auth (401), billing (402), forbidden (403), or not-found (404) response would make image_generate hang for three minutes. Only retry transient statuses (408/409/425/429/5xx); surface everything else immediately. - Add 5 tests covering fail-fast on 401/403/404 and retry on 429/503. * fix(krea): point users at the real API token dashboard URL Three call sites linked users to dashboard pages that don't exist: - hermes_cli/config.py: https://www.krea.ai/app/api/tokens - plugins/image_gen/krea/__init__.py get_setup_schema: https://www.krea.ai/api-keys - plugins/image_gen/krea/__init__.py auth_required error: https://www.krea.ai/api-keys Per Krea's own docs (https://docs.krea.ai/developers/api-keys-and-billing), the real dashboard URL is https://www.krea.ai/settings/api-tokens. All three sites now point there. #AI commit#

…search#33506) fal announced Krea 2 day-0 as an official API partner on 2026-05-27. Add both variants to the FAL_MODELS catalog so they appear in the 'hermes tools' model picker alongside flux-2, gpt-image, nano-banana, etc. Users who already bill through FAL or Nous Portal subscription can now use Krea without registering directly with Krea. Model IDs (as listed in fal's launch announcement): fal-ai/krea/v2/medium/text-to-image — $0.030 / image fal-ai/krea/v2/large/text-to-image — $0.060 / image Both share the same parameter schema: - aspect_ratio (1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 2:3, 9:16) mapped from our 3 abstract ratios via size_style='aspect_ratio' - creativity (raw|low|medium|high; default medium) - seed (reproducibility) - image_style_references (up to 10 per Krea's API spec) No num_inference_steps / guidance_scale / num_images — Krea 2 does not expose those, and the supports-set filter strips them defensively if the agent ever passes them. This is the FAL-routed variant. The separate native-Krea-API plugin shipped in PR NousResearch#33236 (plugins/image_gen/krea/) remains available for users who want to bill directly through Krea's API with their own key. Both routes converge on the same underlying model. Nous Portal managed-FAL gateway: this commit makes the model IDs known to the catalog and the picker. The Portal team will need to allowlist these two endpoint slugs on the fal-queue origin server-side for them to flow through the managed billing path. #AI commit#

…search#33506) fal announced Krea 2 day-0 as an official API partner on 2026-05-27. Add both variants to the FAL_MODELS catalog so they appear in the 'hermes tools' model picker alongside flux-2, gpt-image, nano-banana, etc. Users who already bill through FAL or Nous Portal subscription can now use Krea without registering directly with Krea. Model IDs (as listed in fal's launch announcement): fal-ai/krea/v2/medium/text-to-image — $0.030 / image fal-ai/krea/v2/large/text-to-image — $0.060 / image Both share the same parameter schema: - aspect_ratio (1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 2:3, 9:16) mapped from our 3 abstract ratios via size_style='aspect_ratio' - creativity (raw|low|medium|high; default medium) - seed (reproducibility) - image_style_references (up to 10 per Krea's API spec) No num_inference_steps / guidance_scale / num_images — Krea 2 does not expose those, and the supports-set filter strips them defensively if the agent ever passes them. This is the FAL-routed variant. The separate native-Krea-API plugin shipped in PR NousResearch#33236 (plugins/image_gen/krea/) remains available for users who want to bill directly through Krea's API with their own key. Both routes converge on the same underlying model. Nous Portal managed-FAL gateway: this commit makes the model IDs known to the catalog and the picker. The Portal team will need to allowlist these two endpoint slugs on the fal-queue origin server-side for them to flow through the managed billing path.

…usResearch#33236) * feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large) New built-in image_gen backend wrapping Krea's Krea 2 foundation image model family. Auto-discovered like the other image_gen plugins and appears in 'hermes tools' → Image Generation → Krea. Krea's API is asynchronous — submit returns a job_id, poll /jobs/{id} until terminal. The provider hides that behind the synchronous ImageGenProvider.generate() contract: submit, poll every 2s with light backoff (max 5s), 3-minute ceiling matching Krea's hosted-tool timeout. Result URL is materialised to $HERMES_HOME/cache/images/ to avoid CDN-expiry 404s downstream (same fix as xAI NousResearch#26942). Models: - krea-2-medium (default — Krea's 'start here' recommendation) - krea-2-large Aspect ratios map landscape→16:9, square→1:1, portrait→9:16. Resolution: 1K (Krea's only current option). Kwarg passthrough: seed, creativity (raw/low/medium/high), styles, image_style_references (capped 10), moodboards (capped 1) — matches Krea's per-request limits. Unknown kwargs are ignored. Config knobs (config.yaml): image_gen.provider: krea image_gen.krea.model: krea-2-medium | krea-2-large image_gen.krea.creativity: raw | low | medium | high Env overrides: KREA_API_KEY (required), KREA_IMAGE_MODEL. KREA_API_KEY is registered in OPTIONAL_ENV_VARS so 'hermes setup' prompts for it. 31 new tests; image_gen suite + picker + tools_config: 211/211. * fix(image_gen/krea): address review feedback - Update KREA_API_KEY setup URL to the canonical token-creation page (https://www.krea.ai/app/api/tokens). The previous URL returned 404. - Fail fast on non-retryable HTTP statuses during poll. The previous loop retried every HTTPError for the full 180s deadline, so an auth (401), billing (402), forbidden (403), or not-found (404) response would make image_generate hang for three minutes. Only retry transient statuses (408/409/425/429/5xx); surface everything else immediately. - Add 5 tests covering fail-fast on 401/403/404 and retry on 429/503. * fix(krea): point users at the real API token dashboard URL Three call sites linked users to dashboard pages that don't exist: - hermes_cli/config.py: https://www.krea.ai/app/api/tokens - plugins/image_gen/krea/__init__.py get_setup_schema: https://www.krea.ai/api-keys - plugins/image_gen/krea/__init__.py auth_required error: https://www.krea.ai/api-keys Per Krea's own docs (https://docs.krea.ai/developers/api-keys-and-billing), the real dashboard URL is https://www.krea.ai/settings/api-tokens. All three sites now point there.

…search#33506) fal announced Krea 2 day-0 as an official API partner on 2026-05-27. Add both variants to the FAL_MODELS catalog so they appear in the 'hermes tools' model picker alongside flux-2, gpt-image, nano-banana, etc. Users who already bill through FAL or Nous Portal subscription can now use Krea without registering directly with Krea. Model IDs (as listed in fal's launch announcement): fal-ai/krea/v2/medium/text-to-image — $0.030 / image fal-ai/krea/v2/large/text-to-image — $0.060 / image Both share the same parameter schema: - aspect_ratio (1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 2:3, 9:16) mapped from our 3 abstract ratios via size_style='aspect_ratio' - creativity (raw|low|medium|high; default medium) - seed (reproducibility) - image_style_references (up to 10 per Krea's API spec) No num_inference_steps / guidance_scale / num_images — Krea 2 does not expose those, and the supports-set filter strips them defensively if the agent ever passes them. This is the FAL-routed variant. The separate native-Krea-API plugin shipped in PR NousResearch#33236 (plugins/image_gen/krea/) remains available for users who want to bill directly through Krea's API with their own key. Both routes converge on the same underlying model. Nous Portal managed-FAL gateway: this commit makes the model IDs known to the catalog and the picker. The Portal team will need to allowlist these two endpoint slugs on the fal-queue origin server-side for them to flow through the managed billing path.

…usResearch#33236) * feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large) New built-in image_gen backend wrapping Krea's Krea 2 foundation image model family. Auto-discovered like the other image_gen plugins and appears in 'hermes tools' → Image Generation → Krea. Krea's API is asynchronous — submit returns a job_id, poll /jobs/{id} until terminal. The provider hides that behind the synchronous ImageGenProvider.generate() contract: submit, poll every 2s with light backoff (max 5s), 3-minute ceiling matching Krea's hosted-tool timeout. Result URL is materialised to $HERMES_HOME/cache/images/ to avoid CDN-expiry 404s downstream (same fix as xAI NousResearch#26942). Models: - krea-2-medium (default — Krea's 'start here' recommendation) - krea-2-large Aspect ratios map landscape→16:9, square→1:1, portrait→9:16. Resolution: 1K (Krea's only current option). Kwarg passthrough: seed, creativity (raw/low/medium/high), styles, image_style_references (capped 10), moodboards (capped 1) — matches Krea's per-request limits. Unknown kwargs are ignored. Config knobs (config.yaml): image_gen.provider: krea image_gen.krea.model: krea-2-medium | krea-2-large image_gen.krea.creativity: raw | low | medium | high Env overrides: KREA_API_KEY (required), KREA_IMAGE_MODEL. KREA_API_KEY is registered in OPTIONAL_ENV_VARS so 'hermes setup' prompts for it. 31 new tests; image_gen suite + picker + tools_config: 211/211. * fix(image_gen/krea): address review feedback - Update KREA_API_KEY setup URL to the canonical token-creation page (https://www.krea.ai/app/api/tokens). The previous URL returned 404. - Fail fast on non-retryable HTTP statuses during poll. The previous loop retried every HTTPError for the full 180s deadline, so an auth (401), billing (402), forbidden (403), or not-found (404) response would make image_generate hang for three minutes. Only retry transient statuses (408/409/425/429/5xx); surface everything else immediately. - Add 5 tests covering fail-fast on 401/403/404 and retry on 429/503. * fix(krea): point users at the real API token dashboard URL Three call sites linked users to dashboard pages that don't exist: - hermes_cli/config.py: https://www.krea.ai/app/api/tokens - plugins/image_gen/krea/__init__.py get_setup_schema: https://www.krea.ai/api-keys - plugins/image_gen/krea/__init__.py auth_required error: https://www.krea.ai/api-keys Per Krea's own docs (https://docs.krea.ai/developers/api-keys-and-billing), the real dashboard URL is https://www.krea.ai/settings/api-tokens. All three sites now point there.

…search#33506) fal announced Krea 2 day-0 as an official API partner on 2026-05-27. Add both variants to the FAL_MODELS catalog so they appear in the 'hermes tools' model picker alongside flux-2, gpt-image, nano-banana, etc. Users who already bill through FAL or Nous Portal subscription can now use Krea without registering directly with Krea. Model IDs (as listed in fal's launch announcement): fal-ai/krea/v2/medium/text-to-image — $0.030 / image fal-ai/krea/v2/large/text-to-image — $0.060 / image Both share the same parameter schema: - aspect_ratio (1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 2:3, 9:16) mapped from our 3 abstract ratios via size_style='aspect_ratio' - creativity (raw|low|medium|high; default medium) - seed (reproducibility) - image_style_references (up to 10 per Krea's API spec) No num_inference_steps / guidance_scale / num_images — Krea 2 does not expose those, and the supports-set filter strips them defensively if the agent ever passes them. This is the FAL-routed variant. The separate native-Krea-API plugin shipped in PR NousResearch#33236 (plugins/image_gen/krea/) remains available for users who want to bill directly through Krea's API with their own key. Both routes converge on the same underlying model. Nous Portal managed-FAL gateway: this commit makes the model IDs known to the catalog and the picker. The Portal team will need to allowlist these two endpoint slugs on the fal-queue origin server-side for them to flow through the managed billing path.

…usResearch#33236) * feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large) New built-in image_gen backend wrapping Krea's Krea 2 foundation image model family. Auto-discovered like the other image_gen plugins and appears in 'hermes tools' → Image Generation → Krea. Krea's API is asynchronous — submit returns a job_id, poll /jobs/{id} until terminal. The provider hides that behind the synchronous ImageGenProvider.generate() contract: submit, poll every 2s with light backoff (max 5s), 3-minute ceiling matching Krea's hosted-tool timeout. Result URL is materialised to $HERMES_HOME/cache/images/ to avoid CDN-expiry 404s downstream (same fix as xAI NousResearch#26942). Models: - krea-2-medium (default — Krea's 'start here' recommendation) - krea-2-large Aspect ratios map landscape→16:9, square→1:1, portrait→9:16. Resolution: 1K (Krea's only current option). Kwarg passthrough: seed, creativity (raw/low/medium/high), styles, image_style_references (capped 10), moodboards (capped 1) — matches Krea's per-request limits. Unknown kwargs are ignored. Config knobs (config.yaml): image_gen.provider: krea image_gen.krea.model: krea-2-medium | krea-2-large image_gen.krea.creativity: raw | low | medium | high Env overrides: KREA_API_KEY (required), KREA_IMAGE_MODEL. KREA_API_KEY is registered in OPTIONAL_ENV_VARS so 'hermes setup' prompts for it. 31 new tests; image_gen suite + picker + tools_config: 211/211. * fix(image_gen/krea): address review feedback - Update KREA_API_KEY setup URL to the canonical token-creation page (https://www.krea.ai/app/api/tokens). The previous URL returned 404. - Fail fast on non-retryable HTTP statuses during poll. The previous loop retried every HTTPError for the full 180s deadline, so an auth (401), billing (402), forbidden (403), or not-found (404) response would make image_generate hang for three minutes. Only retry transient statuses (408/409/425/429/5xx); surface everything else immediately. - Add 5 tests covering fail-fast on 401/403/404 and retry on 429/503. * fix(krea): point users at the real API token dashboard URL Three call sites linked users to dashboard pages that don't exist: - hermes_cli/config.py: https://www.krea.ai/app/api/tokens - plugins/image_gen/krea/__init__.py get_setup_schema: https://www.krea.ai/api-keys - plugins/image_gen/krea/__init__.py auth_required error: https://www.krea.ai/api-keys Per Krea's own docs (https://docs.krea.ai/developers/api-keys-and-billing), the real dashboard URL is https://www.krea.ai/settings/api-tokens. All three sites now point there.

…search#33506) fal announced Krea 2 day-0 as an official API partner on 2026-05-27. Add both variants to the FAL_MODELS catalog so they appear in the 'hermes tools' model picker alongside flux-2, gpt-image, nano-banana, etc. Users who already bill through FAL or Nous Portal subscription can now use Krea without registering directly with Krea. Model IDs (as listed in fal's launch announcement): fal-ai/krea/v2/medium/text-to-image — $0.030 / image fal-ai/krea/v2/large/text-to-image — $0.060 / image Both share the same parameter schema: - aspect_ratio (1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 2:3, 9:16) mapped from our 3 abstract ratios via size_style='aspect_ratio' - creativity (raw|low|medium|high; default medium) - seed (reproducibility) - image_style_references (up to 10 per Krea's API spec) No num_inference_steps / guidance_scale / num_images — Krea 2 does not expose those, and the supports-set filter strips them defensively if the agent ever passes them. This is the FAL-routed variant. The separate native-Krea-API plugin shipped in PR NousResearch#33236 (plugins/image_gen/krea/) remains available for users who want to bill directly through Krea's API with their own key. Both routes converge on the same underlying model. Nous Portal managed-FAL gateway: this commit makes the model IDs known to the catalog and the picker. The Portal team will need to allowlist these two endpoint slugs on the fal-queue origin server-side for them to flow through the managed billing path.

alt-glitch added type/feature New feature or request comp/plugins Plugin system and bundled plugins tool/vision Vision analysis and image generation P3 Low — cosmetic, nice to have labels May 27, 2026

albertsalgueda reviewed May 27, 2026

View reviewed changes

teknium1 added 2 commits May 27, 2026 10:42

teknium1 merged commit 9919caf into main May 27, 2026
25 checks passed

teknium1 deleted the hermes/hermes-187cc3e8 branch May 27, 2026 18:01

r266-tech mentioned this pull request May 27, 2026

docs(built-in-plugins): add image_gen/krea row and fix xai model id #33404

Open

teknium1 mentioned this pull request May 27, 2026

feat(image_gen/fal): add Krea 2 Medium + Large to FAL catalog #33506

Merged

BrewTestBot mentioned this pull request May 28, 2026

hermes-agent 2026.5.28 Homebrew/homebrew-core#285115

Merged

1 task

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large)#33236

feat(image_gen): add Krea provider plugin (Krea 2 Medium + Large)#33236
teknium1 merged 3 commits into
mainfrom
hermes/hermes-187cc3e8

teknium1 commented May 27, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented May 27, 2026 •

edited

Loading

Uh oh!

albertsalgueda May 27, 2026 •

edited

Loading

Uh oh!

teknium1 May 27, 2026

Uh oh!

albertsalgueda May 27, 2026

Uh oh!

teknium1 May 27, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

teknium1 commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

Design notes

Config

Validation

Infographic

Uh oh!

github-actions Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔎 Lint report: hermes/hermes-187cc3e8 vs origin/main

ruff

ty (type checker)

Uh oh!

albertsalgueda May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

teknium1 May 27, 2026

Choose a reason for hiding this comment

Uh oh!

albertsalgueda May 27, 2026

Choose a reason for hiding this comment

Uh oh!

teknium1 May 27, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

teknium1 commented May 27, 2026 •

edited

Loading

github-actions Bot commented May 27, 2026 •

edited

Loading

🔎 Lint report: `hermes/hermes-187cc3e8` vs `origin/main`

albertsalgueda May 27, 2026 •

edited

Loading