feat: direct api key and cheap model by ilblackdragon · Pull Request #116 · nearai/ironclaw

ilblackdragon · 2026-02-17T01:20:41Z

Fixes from #20

Allow using IronClaw with any OpenAI-compatible API provider (e.g. Anthropic Claude) via API key, without requiring NEAR AI session auth. Changes: - Skip session authentication in chat_completions mode (API key auth) - Skip first-run onboard check when NEARAI_API_KEY is configured - Add `cheap_model` config field (NEARAI_CHEAP_MODEL env var) for a secondary lightweight model used for heartbeat, routing, evaluation - Add `create_cheap_llm_provider()` factory in llm module - Add `cheap_llm` to AgentDeps with fallback to main model - Route heartbeat through cheap model to reduce costs - Fix wizard compilation for new config field Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

- Check API key presence (not api_mode) for auth skip (ilblackdragon) - Add Settings::load() call in check_onboard_needed (ilblackdragon) - Warn and ignore cheap_model for non-NearAi backends (ilblackdragon) - Add unit tests for create_cheap_llm_provider (ilblackdragon) - Minor formatting cleanup in cheap provider match arm Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

gemini-code-assist · 2026-02-17T01:20:57Z

Summary of Changes

Hello @ilblackdragon, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request enhances the system's flexibility and cost-efficiency by introducing the capability to use a dedicated, less resource-intensive LLM for background and auxiliary functions. It also improves the user experience by making the authentication process more intelligent, recognizing direct API key usage and adjusting onboarding requirements accordingly. These changes aim to optimize operational costs and simplify initial setup for certain configurations.

Highlights

Cheap LLM Integration: Introduced an optional 'cheap LLM' provider for lightweight tasks such as heartbeats, routing, and evaluation, allowing for cost optimization by using a less expensive model for non-critical operations. This cheap LLM falls back to the main LLM if not explicitly configured.
Configuration for Cheap Model: Added a cheap_model field to NearAiConfig which can be set via the NEARAI_CHEAP_MODEL environment variable, enabling users to specify a distinct model for these lightweight tasks.
Refined Authentication Logic: Modified the authentication flow for NEAR AI backends to bypass session-based authentication when a direct API key is provided, streamlining the setup process for users with API keys.
Onboarding Experience Improvement: Updated the onboarding check to consider the presence of the NEARAI_API_KEY environment variable, preventing unnecessary onboarding prompts when an API key is already configured.

Changelog

src/agent/agent_loop.rs
- Added cheap_llm field to AgentDeps struct to hold an optional cheap LLM provider.
- Implemented a cheap_llm() method in Agent to retrieve the cheap LLM, falling back to the main LLM if not set.
- Updated the spawn_heartbeat function to utilize the newly introduced cheap_llm for heartbeat operations.
src/config.rs
- Introduced an optional cheap_model field to NearAiConfig for specifying a lightweight model.
- Configured the cheap_model to be loaded from the NEARAI_CHEAP_MODEL environment variable.
src/llm/mod.rs
- Added a new public function create_cheap_llm_provider responsible for initializing the cheap LLM based on configuration.
- Included logic within create_cheap_llm_provider to warn and return None if NEARAI_CHEAP_MODEL is set but the backend is not NearAi.
- Provided unit tests for create_cheap_llm_provider to cover cases where it's not configured, configured correctly, and ignored for non-NearAi backends.
src/main.rs
- Imported the new create_cheap_llm_provider function.
- Modified the session authentication check to only proceed if the backend is NearAi AND no API key is present.
- Initialized the cheap_llm provider during application startup and passed it to AgentDeps.
- Updated the check_onboard_needed function to also check for the NEARAI_API_KEY environment variable, influencing onboarding decisions.
src/setup/wizard.rs
- Initialized the cheap_model field to None within the NearAiConfig structure during the setup wizard's configuration generation.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	`/gemini review`	Performs a code review for the current pull request in its current state.
Pull Request Summary	`/gemini summary`	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	`/gemini help`	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

gemini-code-assist

Code Review

The pull request adds support for direct API key authentication and a secondary 'cheap' LLM provider for lightweight tasks. The implementation is generally sound and integrates well with the existing configuration and agent loop. I have identified a few areas for improvement regarding authentication logic robustness and code duplication.

gemini-code-assist · 2026-02-17T01:24:04Z

+    match cheap_config.api_mode {
+        NearAiApiMode::Responses => Ok(Some(Arc::new(NearAiProvider::new(cheap_config, session)))),
+        NearAiApiMode::ChatCompletions => {
+            Ok(Some(Arc::new(NearAiChatProvider::new(cheap_config)?)))
+        }
+    }


The logic for creating the provider based on api_mode is identical to create_llm_provider_with_config. You can refactor this to use the existing helper function to reduce duplication and ensure consistent provider initialization.

gemini-code-assist · 2026-02-17T01:24:04Z

+    // Session-based auth is only needed for NEAR AI backend without an API key.
+    // ChatCompletions mode with an API key skips session auth entirely.
+    if config.llm.backend == ironclaw::config::LlmBackend::NearAi
+        && config.llm.nearai.api_key.is_none()
+    {
        session.ensure_authenticated().await?;
    }


The condition for skipping session-based authentication should ideally check the api_mode rather than just the presence of an API key. While api_key.is_some() usually implies ChatCompletions mode, a user could explicitly configure Responses mode while having an API key set. In Responses mode, the NearAiProvider always requires a session token, so ensure_authenticated() must be called.

Copilot

Pull request overview

This PR adds support for direct API key authentication with NEAR AI and introduces a cheap model configuration for cost-sensitive operations. It builds upon PR #20 by implementing the infrastructure for using a secondary, lightweight LLM for tasks like heartbeat monitoring, with future support planned for routing and evaluation.

Changes:

Added NEARAI_CHEAP_MODEL environment variable to configure a secondary lightweight model for cost-sensitive tasks
Modified session authentication logic to skip OAuth when using ChatCompletions mode with an API key
Enhanced first-run detection to skip onboarding when an API key is already configured

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/config.rs	Added `cheap_model` optional field to `NearAiConfig` structure and loaded from environment variable
src/llm/mod.rs	Implemented `create_cheap_llm_provider()` factory function with comprehensive unit tests
src/main.rs	Integrated cheap LLM provider initialization, updated session auth logic to skip when API key present, modified onboarding check
src/agent/agent_loop.rs	Added `cheap_llm` field to `AgentDeps`, implemented `cheap_llm()` accessor with fallback, routed heartbeat to use cheap model
src/setup/wizard.rs	Added `cheap_model: None` to fix NearAiConfig struct initialization

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-17T01:26:08Z

+    // Initialize cheap LLM provider for lightweight tasks (heartbeat, evaluation)
+    let cheap_llm = create_cheap_llm_provider(&config.llm, session.clone())?;
+    if let Some(ref cheap) = cheap_llm {
+        tracing::info!("Cheap LLM provider initialized: {}", cheap.model_name());
+    }


Consider adding a warning when the cheap model is the same as the main model, similar to the fallback model check above (lines 522-526). This would help users avoid misconfiguration where they set NEARAI_CHEAP_MODEL to the same value as NEARAI_MODEL, defeating the purpose of having a cheaper model for lightweight tasks.

Copilot · 2026-02-17T01:26:08Z

+
+        assert!(result.is_ok());
+        assert!(result.unwrap().is_none());
+    }


Consider adding a test case for creating a cheap LLM provider with ChatCompletions API mode. The existing tests only cover the Responses mode, but the code supports both modes (lines 213-218). This would improve test coverage for this feature.

Suggested change

}

}

#[test]

fn test_create_cheap_llm_provider_with_chat_completions_mode() {

let mut config = test_llm_config();

config.nearai.cheap_model = Some("cheap-test-model".to_string());

config.nearai.api_mode = NearAiApiMode::ChatCompletions;

let session = Arc::new(SessionManager::new(SessionConfig::default()));

let result = create_cheap_llm_provider(&config, session);

assert!(result.is_ok());

let provider = result.unwrap();

assert!(provider.is_some());

assert_eq!(provider.unwrap().model_name(), "cheap-test-model");

}

* feat: Support direct API key auth and cheap model routing Allow using IronClaw with any OpenAI-compatible API provider (e.g. Anthropic Claude) via API key, without requiring NEAR AI session auth. Changes: - Skip session authentication in chat_completions mode (API key auth) - Skip first-run onboard check when NEARAI_API_KEY is configured - Add `cheap_model` config field (NEARAI_CHEAP_MODEL env var) for a secondary lightweight model used for heartbeat, routing, evaluation - Add `create_cheap_llm_provider()` factory in llm module - Add `cheap_llm` to AgentDeps with fallback to main model - Route heartbeat through cheap model to reduce costs - Fix wizard compilation for new config field Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: address PR nearai#20 review feedback - Check API key presence (not api_mode) for auth skip (ilblackdragon) - Add Settings::load() call in check_onboard_needed (ilblackdragon) - Warn and ignore cheap_model for non-NearAi backends (ilblackdragon) - Add unit tests for create_cheap_llm_provider (ilblackdragon) - Minor formatting cleanup in cheap provider match arm Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Samuel Barbosa <sambarbosaa@gmail.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

desamtralized and others added 2 commits February 16, 2026 16:49

Copilot AI review requested due to automatic review settings February 17, 2026 01:20

Copilot started reviewing on behalf of ilblackdragon February 17, 2026 01:21 View session

gemini-code-assist Bot reviewed Feb 17, 2026

View reviewed changes

ilblackdragon merged commit 72623c9 into main Feb 17, 2026
6 checks passed

ilblackdragon deleted the feat/direct-api-key-and-cheap-model branch February 17, 2026 01:24

ilblackdragon mentioned this pull request Feb 17, 2026

feat: Support direct API key auth and cheap model routing #20

Closed

5 tasks

Copilot AI reviewed Feb 17, 2026

View reviewed changes

github-actions Bot mentioned this pull request Feb 17, 2026

chore: release v0.3.0 #117

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: direct api key and cheap model#116

feat: direct api key and cheap model#116
ilblackdragon merged 2 commits intomainfrom
feat/direct-api-key-and-cheap-model

ilblackdragon commented Feb 17, 2026 •

edited

Loading

Uh oh!

gemini-code-assist Bot commented Feb 17, 2026

Uh oh!

gemini-code-assist Bot left a comment

Uh oh!

gemini-code-assist Bot Feb 17, 2026

Uh oh!

gemini-code-assist Bot Feb 17, 2026

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 17, 2026

Uh oh!

Copilot AI Feb 17, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

-    }
+    }
+    #[test]
+    fn test_create_cheap_llm_provider_with_chat_completions_mode() {
+        let mut config = test_llm_config();
+        config.nearai.cheap_model = Some("cheap-test-model".to_string());
+        config.nearai.api_mode = NearAiApiMode::ChatCompletions;
+        let session = Arc::new(SessionManager::new(SessionConfig::default()));
+        let result = create_cheap_llm_provider(&config, session);
+        assert!(result.is_ok());
+        let provider = result.unwrap();
+        assert!(provider.is_some());
+        assert_eq!(provider.unwrap().model_name(), "cheap-test-model");
+    }

Conversation

ilblackdragon commented Feb 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gemini-code-assist Bot commented Feb 17, 2026

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist Bot Feb 17, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Feb 17, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 17, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 17, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ilblackdragon commented Feb 17, 2026 •

edited

Loading