docs: add Eden AI integration guides for Bifrost and Lynkr (AI Gateway)#70
docs: add Eden AI integration guides for Bifrost and Lynkr (AI Gateway)#70MVS-source wants to merge 1 commit into
Conversation
Add two new integration guides under a new "AI Gateway" nav group: - Bifrost: OpenAI-compatible LLM gateway (Custom Provider config) - Lynkr: tier-based routing gateway (built-in Eden AI provider) Both point at https://api.edenai.run/v3. Env var follows the docs house style (EDEN_AI_API_KEY) for Bifrost; Lynkr uses its derived EDENAI_API_KEY (<PROVIDER>_API_KEY convention). Includes square icon SVGs and docs.json nav entries. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
WalkthroughAdded Bifrost and Lynkr integration documentation, including setup, configuration, routing, models, environment variables, metadata, and navigation under a new V3 AI Gateway group. ChangesAI Gateway documentation
Estimated code review effort: 2 (Simple) | ~10 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
|
Preview deployment for your docs. Learn more about Mintlify Previews.
💡 Tip: Enable Workflows to automatically generate PRs for you. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
v3/integrations/bifrost.mdx (1)
107-107: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low valueUse standardized token placeholders.
As per coding guidelines, "Distinguish token types in code examples:
api_tokenfor production andsandbox_api_tokenfor testing (no real provider calls)". Please update these environment variables to use the standardized placeholder.
v3/integrations/bifrost.mdx#L107-L107: Replaceyour_api_key_herewithapi_token.v3/integrations/lynkr.mdx#L51-L51: Replaceyour-edenai-keywithapi_token.v3/integrations/lynkr.mdx#L72-L72: Replaceyour-edenai-keywithapi_token.v3/integrations/lynkr.mdx#L124-L124: Replaceyour_api_key_herewithapi_token.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@v3/integrations/bifrost.mdx` at line 107, Standardize the token placeholders in v3/integrations/bifrost.mdx lines 107-107 and v3/integrations/lynkr.mdx lines 51-51, 72-72, and 124-124 by replacing each provider-specific or generic placeholder with api_token; no other changes are needed.Source: Coding guidelines
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@v3/integrations/lynkr.mdx`:
- Around line 93-96: Update the Lynkr environment variable definitions so the
existing TIER_HARD and TIER_EXPERT entries are renamed to TIER_COMPLEX and
TIER_REASONING, respectively, while preserving their current model assignments
and the TIER_SIMPLE and TIER_MEDIUM entries.
---
Nitpick comments:
In `@v3/integrations/bifrost.mdx`:
- Line 107: Standardize the token placeholders in v3/integrations/bifrost.mdx
lines 107-107 and v3/integrations/lynkr.mdx lines 51-51, 72-72, and 124-124 by
replacing each provider-specific or generic placeholder with api_token; no other
changes are needed.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 3f32cdb5-14a9-4a55-adcb-c0476326e739
⛔ Files ignored due to path filters (2)
integration-logo/bifrost-icondoc.svgis excluded by!**/*.svgintegration-logo/lynkr-icondoc.svgis excluded by!**/*.svg
📒 Files selected for processing (3)
docs.jsonv3/integrations/bifrost.mdxv3/integrations/lynkr.mdx
| export TIER_SIMPLE=edenai:mistral/mistral-small-latest | ||
| export TIER_MEDIUM=edenai:openai/gpt-4o-mini | ||
| export TIER_HARD=edenai:anthropic/claude-sonnet-4-5 | ||
| export TIER_EXPERT=edenai:openai/gpt-4o |
There was a problem hiding this comment.
📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
🧩 Analysis chain
🌐 Web query:
What are the tier-based routing environment variables (e.g., TIER_SIMPLE, TIER_MEDIUM) used in the Lynkr LLM gateway?
💡 Result:
In the Lynkr LLM gateway, tier-based routing is configured using four mandatory environment variables [1][2]. When all four are defined, Lynkr scores incoming requests for complexity (on a scale of 0-100) and routes them to the specific provider and model assigned to that tier [1][3]. The four required tier-based routing environment variables are: - TIER_SIMPLE: Used for simple tasks (complexity score 0-25) [1][4]. - TIER_MEDIUM: Used for medium-complexity tasks (complexity score 26-50) [1][4]. - TIER_COMPLEX: Used for complex tasks (complexity score 51-75) [1][4]. - TIER_REASONING: Used for highly complex or reasoning-intensive tasks (complexity score 76-100) [1][4]. These variables must be set in the format provider:model (e.g., TIER_SIMPLE=ollama:llama3.2) [1][2]. Important configuration details: - Activation: All four variables must be set simultaneously to enable tiered routing [1][5]. If any variable is missing, tiered routing is disabled, and the system defaults to the provider specified in the MODEL_PROVIDER variable [3][5]. - Priority: When all four are correctly configured, they override the MODEL_PROVIDER for routing decisions [2][4]. - Fallback: The MODEL_PROVIDER variable must still be set, as it is used for system startup checks and as a default provider in specific edge cases [2][5]. Lynkr supports various providers, including ollama, openai, azure-openai, openrouter, databricks, bedrock, vertex, zai, moonshot, llamacpp, and lmstudio [6][2].
Citations:
- 1: https://github.com/Fast-Editor/Lynkr/blob/main/documentation/routing.md
- 2: https://github.com/Fast-Editor/Lynkr/blob/cba14ec5/documentation/providers.md
- 3: https://github.com/Fast-Editor/Lynkr/blob/cba14ec5/documentation/routing.md
- 4: https://github.com/Fast-Editor/Lynkr/blob/HEAD/documentation/installation.md
- 5: https://github.com/Fast-Editor/Lynkr/blob/main/documentation/faq.md
- 6: https://github.com/Fast-Editor/Lynkr/blob/cba14ec5/.env.example
🏁 Script executed:
#!/bin/bash
set -euo pipefail
printf '--- file outline ---\n'
ast-grep outline v3/integrations/lynkr.mdx --view expanded || true
printf '\n--- relevant file excerpt ---\n'
sed -n '1,180p' v3/integrations/lynkr.mdx
printf '\n--- search for tier env vars in repo ---\n'
rg -n "TIER_(SIMPLE|MEDIUM|COMPLEX|REASONING|HARD|EXPERT)|MODEL_PROVIDER|routing" v3 documentation . -g '!**/node_modules/**' || trueRepository: edenai/docs
Length of output: 240
🌐 Web query:
Lynkr routing documentation TIER_SIMPLE TIER_MEDIUM TIER_COMPLEX TIER_REASONING environment variables
💡 Result:
In Lynkr, the TIER_SIMPLE, TIER_MEDIUM, TIER_COMPLEX, and TIER_REASONING environment variables are used to enable and configure tier-based routing [1][2][3]. When all four variables are set, Lynkr scores incoming requests for complexity (0–100) and automatically routes them to the provider and model specified for the matching tier [1][2][3]. Key details for these variables include: Configuration Requirements: - Format: You must use the provider:model format (e.g., TIER_SIMPLE=ollama:llama3.2) [1][3][4]. - Activation: All four variables must be defined to activate tier-based routing [2][5][3]. If any variable is missing, tier-based routing is disabled, and Lynkr defaults to using the provider defined in the MODEL_PROVIDER environment variable [2][5][3]. - Precedence: When active, these variables override the standard MODEL_PROVIDER setting for routing decisions [2][5][6]. Tier Complexity Mapping: Requests are scored on a scale of 0–100 and mapped to the tiers as follows [2][6]: - TIER_SIMPLE (0–25) - TIER_MEDIUM (26–50) - TIER_COMPLEX (51–75) - TIER_REASONING (76–100) Additional Considerations: - MODEL_PROVIDER: Even when tier-based routing is active, the MODEL_PROVIDER variable remains required for internal processes, such as startup checks, provider discovery, and as a fallback if a tier variable lacks a provider prefix [5][3]. - Fallback: You can also configure a FALLBACK_PROVIDER to handle cases where a chosen tier provider fails [2][6][7]. - Setup: It is recommended to configure these in your.env file using plain KEY=VALUE syntax without an export prefix [8][4].
Citations:
- 1: https://github.com/Fast-Editor/Lynkr/blob/main/documentation/routing.md
- 2: https://github.com/Fast-Editor/Lynkr/blob/HEAD/documentation/installation.md
- 3: https://github.com/Fast-Editor/Lynkr/blob/cba14ec5/documentation/routing.md
- 4: https://github.com/Fast-Editor/Lynkr/blob/cba14ec5/.env.example
- 5: https://github.com/Fast-Editor/Lynkr/blob/cba14ec5/documentation/providers.md
- 6: https://github.com/Fast-Editor/Lynkr/blob/main/documentation/installation.md
- 7: https://github.com/Fast-Editor/Lynkr/blob/main/documentation/docker.md
- 8: https://github.com/Fast-Editor/Lynkr/blob/main/.env.example
Use Lynkr’s documented tier names. TIER_HARD and TIER_EXPERT should be TIER_COMPLEX and TIER_REASONING; Lynkr expects TIER_SIMPLE, TIER_MEDIUM, TIER_COMPLEX, and TIER_REASONING.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@v3/integrations/lynkr.mdx` around lines 93 - 96, Update the Lynkr environment
variable definitions so the existing TIER_HARD and TIER_EXPERT entries are
renamed to TIER_COMPLEX and TIER_REASONING, respectively, while preserving their
current model assignments and the TIER_SIMPLE and TIER_MEDIUM entries.
|
Preview deployment for your docs. Learn more about Mintlify Previews.
💡 Tip: Enable Workflows to automatically generate PRs for you. |
|
Remplacée par #73 (historique de commits nettoyé, même contenu). |
2 participants
What
Two new integration guides, grouped under a new AI Gateway nav section:
config.json, no fork/buildBoth route to Eden AI's OpenAI-compatible V3 API (
https://api.edenai.run/v3) for access to 500+ models behind one key.Files
v3/integrations/bifrost.mdx(new)v3/integrations/lynkr.mdx(new)integration-logo/bifrost-icondoc.svg,lynkr-icondoc.svg(new, square marks)docs.json— new AI Gateway group (iconnetwork-wired) with both pagesConventions
<TechArticleSchema>+<CodeGroup>).EDEN_AI_API_KEY(house style) for Bifrost;EDENAI_API_KEYfor Lynkr since it derives the name from the provider (<PROVIDER>_API_KEY).To confirm before merge
custom_provider_config.allowed_requestskeys (chat_completion/embedding) against the current Bifrost schema.TIER_*variable names and default port (8081).🤖 Generated with Claude Code
Summary by CodeRabbit