Skip to content

feat: add Engy as OpenAI-compatible AI Gateway provider (full chat catalog) - #5838

Open
crabbytt wants to merge 7 commits into
Helicone:mainfrom
crabbytt:cursor/add-engy-provider-58f1
Open

crabbytt wants to merge 7 commits into
Helicone:mainfrom
crabbytt:cursor/add-engy-provider-58f1

Conversation

@crabbytt

@crabbytt crabbytt commented Oct 15, 2026 •

Copy link
Copy Markdown

Ticket

N/A — third-party provider integration (adds Engy as a new AI Gateway provider).

Component/Service

  • Web (Frontend)
  • Jawn (Backend)
  • Worker (Proxy)
  • Bifrost (Marketing)
  • AI Gateway
  • Packages
  • Infrastructure/Docker
  • Documentation

Type of Change

  • Bug fix
  • New feature
  • Breaking change
  • Documentation update
  • Performance improvement
  • Refactoring

Deployment Notes

  • No special deployment steps required

Context

Adds Engy (https://engy.ai/) as an OpenAI-compatible inference provider for the Helicone AI Gateway, with provider registration, a BYOK web card, and endpoint + pricing entries for Engy's full chat catalog (8 models). Implemented following https://docs.helicone.ai/references/provider-integration.

Disclosure: I work on Engy. Happy to iterate on author/model-family placement, ptbEnabled, the logo asset, and any gateway test fixtures maintainers prefer.

Contact: terry@engy.ai

Provider

Field Value
Provider id engy
Display name Engy
API base https://api.engy.ai/v1
Chat Completions URL https://api.engy.ai/v1/chat/completions
Auth api-key (Authorization: Bearer $ENGY_API_KEY)
Site / pricing https://engy.ai/
Docs https://engy.ai/docs
Public model list https://api.engy.ai/v1/models

Pattern matches existing OpenAI-compatible providers (e.g. DeepInfra / Novita / Canopy Wave): BaseProvider subclass, buildUrl → v1/chat/completions, OpenAI usage processor.

Models & pricing (USD)

Helicone pricing uses USD per token (input / output); cacheMultipliers.cachedInput = Engy cache_read ÷ input (both USD/1M).

Gateway key Engy model id Input $/1M Output $/1M Cache read $/1M Context Modalities
kimi-k3:engy kimi-k3 1.95 9.75 0.195 1,113,088 text+image→text
glm-5.3:engy glm-5.3 0.98 3.08 0.18 327,680 text→text
glm-5.3-flash:engy glm-5.3-flash 0.135 0.45 0.027 327,680 text+image→text
deepseek-v4.1-flash:engy deepseek-v4.1-flash 0.03 0.12 0.006 1,048,576 text+image→text
glm-5.2:engy glm-5.2 0.68 1.50 0.18 262,144 text→text
deepseek-v4-flash-0731:engy deepseek-v4-flash-0731 0.045 0.09 0.009 1,048,576 text→text
qwen3.8-27b:engy qwen3.8-27b 0.045 0.32 0.015 1,001,536 text+image→text
qwen3.6-35b-a3b:engy qwen3.6-35b-a3b 0.045 0.30 0.015 208,192 text→text

Video model MiniMax-H3 is intentionally not included (not a Chat Completions model). Prices from https://engy.ai/ and the public /v1/models feed (synced 2026-10-23).

Test plan

  • Typecheck / jest / worker AI Gateway suites pass on the contribution branch
  • Provider resolves URL to https://api.engy.ai/v1/chat/completions with Authorization: Bearer
  • Live BYOK request with a real Engy API key (needs an ENGY_API_KEY)

Extra Notes

  • ptbEnabled: false on all Engy endpoints pending pass-through billing confirmation.
  • Logo asset is a placeholder tile; happy to swap official brand art.
  • Author assignments follow existing taxonomy (moonshotai, zai, deepseek, qwen under alibaba/).

Note

Medium Risk
Touches gateway routing, cost registry, and usage accounting for a new provider; mistakes could misprice or misroute traffic, though changes follow existing OpenAI-compatible provider patterns with broad test coverage.

Overview
Introduces Engy as a new OpenAI-compatible AI Gateway provider (api.engy.ai/v1/chat/completions), including EngyProvider, URL/provider mappings, usage via OpenAIUsageProcessor, priority/default routing helpers, and a BYOK card in the web provider list.

Registers nine Engy endpoints (eight net-new model families) with pricing, context limits, supported parameters, and registry aggregation: Qwen3 (qwen3.8-27b, qwen3.6-35b-a3b), DeepSeek V4 (deepseek-v4.1-flash, deepseek-v4-flash-0731), Kimi K3 (kimi-k3, kimi-k3.1-pro), and GLM-5 (glm-5.3, glm-5.3-flash, glm-5.2). All Engy routes use ptbEnabled: false. Cost registry snapshots bump totals (338 endpoints, 22 providers).

Adds AI Gateway BYOK tests for Engy (chat + tool calls) and test fixtures for Engy API keys in worker setup.

Reviewed by Cursor Bugbot for commit 37a9175. Bugbot is set up for automated code reviews on this repo. Configure here.

@cursor

cursor Bot commented Oct 15, 2026

Copy link
Copy Markdown

Bugbot is paused — your team has an unpaid invoice

Bugbot is paused because your team has an unpaid invoice.

Pay it from the billing page in the Cursor dashboard and Bugbot resumes automatically.

(requestId: 993bf12b-e41e-48b4-bf19-f5088a3cfe89)

@greptile-apps greptile-apps Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Your trial has ended. Reactivate Greptile to resume code reviews.

@vercel

vercel Bot commented Oct 15, 2026

Copy link
Copy Markdown

@cursoragent is attempting to deploy a commit to the Helicone Team on Vercel.

A member of the Team first needs to authorize it.

Add EngyProvider (https://api.engy.ai/v1/chat/completions, Bearer auth)
and wire it into the provider registry, priorities, legacy URL-pattern
mappings, heliconeProviderToModelProviderName/dbProviderToProvider
helpers, and the OpenAI-compatible usage processor.
New model families with :engy endpoints (BYOK, ptbEnabled: false):
- moonshotai/kimi-k3 (1,113,088 ctx, text+image)
- zai/glm-5: glm-5.3, glm-5.3-flash, glm-5.2
- deepseek/deepseek-v4: deepseek-v4.1-flash, deepseek-v4-flash-0731
- alibaba/qwen3: qwen3.8-27b, qwen3.6-35b-a3b

Per-token pricing converted from Engy USD/1M list prices;
cacheMultipliers.cachedInput = cache_read / input. Registry snapshots
regenerated with npx jest -u.
@cursor
cursor Bot force-pushed the cursor/add-engy-provider-58f1 branch from bc7914a to f56daa1 Compare October 19, 2026 10:34
@cursor

cursor Bot commented Oct 19, 2026

Copy link
Copy Markdown

Bugbot is paused — your team has an unpaid invoice

Bugbot is paused because your team has an unpaid invoice.

Pay it from the billing page in the Cursor dashboard and Bugbot resumes automatically.

(requestId: 4b0a1198-f1a3-4051-aa0e-364b50394159)

Match live Engy /v1/models: $0.03 input, $0.12 output, and $0.006 cache read per 1M tokens. cachedInput stays 0.2 because 0.006 / 0.03.

Co-authored-by: crabbytt <crabbytt@users.noreply.github.com>
@cursor

cursor Bot commented Oct 23, 2026

Copy link
Copy Markdown

Bugbot is paused — your team has an unpaid invoice

Bugbot is paused because your team has an unpaid invoice.

Pay it from the billing page in the Cursor dashboard and Bugbot resumes automatically.

(requestId: 7c8ef791-4a19-412e-893f-2c86189e11ed)

Register kimi-k3.1-pro (moonshotai/Kimi-K3.1-Pro) in the kimi-k3 model
family with an Engy gateway endpoint: 1,113,088 token context, text +
image + video input, $1.40/$7.00 per 1M tokens input/output with
cachedInput multiplier 0.1 ($0.14/1M cache read), matching the other
Engy endpoints.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants