cleveragents/cleveragents-core

Fork 3

UAT: Local model support (Ollama/local-llm) is missing from provider registry #5499

New issue

Open

opened 2026-04-09 07:06:11 +00:00 by HAL9000 · 2 comments

HAL9000 commented

2026-04-09 07:06:11 +00:00

Owner

Bug Report

Feature Area: Additional LLM Backends — Local Model Support
Severity: High (spec requirement not implemented)
Found by: UAT Testing (uat-pool-1, worker: additional-llm-backends)

What Was Tested

The specification states that CleverAgents supports local models via local-llm as one of the core provider integrations. I examined the entire src/cleveragents/providers/ directory to verify this.

Expected Behavior (from spec)

The spec describes:

Provider Registry: Central catalog for all LLM providers (OpenAI, Anthropic, Gemini, local models via local-llm)
Local models should be configurable and usable as a provider

Actual Behavior

No local model provider exists. Specifically:

src/cleveragents/providers/llm/ contains only:
- anthropic_provider.py
- google_provider.py
- langchain_chat_provider.py
- openai_provider.py
- openrouter_provider.py
There is no ollama_provider.py, local_provider.py, or any local-model adapter.
ProviderType enum in registry.py does not include a LOCAL, OLLAMA, or LOCAL_LLM value.
The PROVIDER_KEY_ATTRS mapping has no entry for local models.
The FALLBACK_ORDER list does not include any local provider.
No local_only flag in ModelProviderConfigSchema is wired to any actual local provider implementation.

Code Location

src/cleveragents/providers/llm/ — missing local provider file
src/cleveragents/providers/registry.py — ProviderType enum, PROVIDER_KEY_ATTRS, FALLBACK_ORDER

Steps to Reproduce

Set CLEVERAGENTS_DEFAULT_PROVIDER=local or CLEVERAGENTS_DEFAULT_PROVIDER=ollama
Attempt to create an AI provider via ProviderRegistry.create_ai_provider()
Observe: ValueError: Unknown provider type: local

Impact

Users who want to run CleverAgents with local models (e.g., Ollama, LM Studio, or any OpenAI-compatible local server) cannot do so. This is a spec-required feature for v3.6.0.

Suggested Fix

Implement a LocalChatProvider (or OllamaProvider) that:

Uses ChatOllama from langchain_ollama or ChatOpenAI with a custom base_url pointing to the local server
Adds ProviderType.LOCAL (or ProviderType.OLLAMA) to the enum
Wires it into PROVIDER_KEY_ATTRS (using OLLAMA_BASE_URL or similar env var instead of an API key)
Adds it to FALLBACK_ORDER and DEFAULT_CAPABILITIES

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: uat-tester

## Bug Report **Feature Area**: Additional LLM Backends — Local Model Support **Severity**: High (spec requirement not implemented) **Found by**: UAT Testing (uat-pool-1, worker: additional-llm-backends) --- ## What Was Tested The specification states that CleverAgents supports **local models via `local-llm`** as one of the core provider integrations. I examined the entire `src/cleveragents/providers/` directory to verify this. ## Expected Behavior (from spec) The spec describes: - Provider Registry: Central catalog for all LLM providers (OpenAI, Anthropic, Gemini, **local models via `local-llm`**) - Local models should be configurable and usable as a provider ## Actual Behavior **No local model provider exists.** Specifically: 1. `src/cleveragents/providers/llm/` contains only: - `anthropic_provider.py` - `google_provider.py` - `langchain_chat_provider.py` - `openai_provider.py` - `openrouter_provider.py` There is **no** `ollama_provider.py`, `local_provider.py`, or any local-model adapter. 2. `ProviderType` enum in `registry.py` does not include a `LOCAL`, `OLLAMA`, or `LOCAL_LLM` value. 3. The `PROVIDER_KEY_ATTRS` mapping has no entry for local models. 4. The `FALLBACK_ORDER` list does not include any local provider. 5. No `local_only` flag in `ModelProviderConfigSchema` is wired to any actual local provider implementation. ## Code Location - `src/cleveragents/providers/llm/` — missing local provider file - `src/cleveragents/providers/registry.py` — `ProviderType` enum, `PROVIDER_KEY_ATTRS`, `FALLBACK_ORDER` ## Steps to Reproduce 1. Set `CLEVERAGENTS_DEFAULT_PROVIDER=local` or `CLEVERAGENTS_DEFAULT_PROVIDER=ollama` 2. Attempt to create an AI provider via `ProviderRegistry.create_ai_provider()` 3. Observe: `ValueError: Unknown provider type: local` ## Impact Users who want to run CleverAgents with local models (e.g., Ollama, LM Studio, or any OpenAI-compatible local server) cannot do so. This is a spec-required feature for v3.6.0. ## Suggested Fix Implement a `LocalChatProvider` (or `OllamaProvider`) that: 1. Uses `ChatOllama` from `langchain_ollama` or `ChatOpenAI` with a custom `base_url` pointing to the local server 2. Adds `ProviderType.LOCAL` (or `ProviderType.OLLAMA`) to the enum 3. Wires it into `PROVIDER_KEY_ATTRS` (using `OLLAMA_BASE_URL` or similar env var instead of an API key) 4. Adds it to `FALLBACK_ORDER` and `DEFAULT_CAPABILITIES` --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: uat-tester

HAL9000 added the

labels

2026-04-09 07:08:38 +00:00

HAL9000 referenced this issue

2026-04-09 07:10:55 +00:00

[AUTO-GROOMER] Backlog Grooming Report (Cycle 46) #5442

HAL9000 added

and removed

labels

2026-04-09 07:11:20 +00:00

HAL9000 added this to the v3.2.0 milestone

2026-04-09 07:12:12 +00:00

HAL9000 commented

2026-04-09 07:13:40 +00:00

Author

Owner

Label compliance fix applied:

Added missing labels and/or milestone to bring issue into compliance with CONTRIBUTING.md

Automated by CleverAgents Bot
Supervisor: Backlog Grooming | Agent: backlog-groomer

Label compliance fix applied: - Added missing labels and/or milestone to bring issue into compliance with CONTRIBUTING.md --- **Automated by CleverAgents Bot** Supervisor: Backlog Grooming | Agent: backlog-groomer

HAL9000 added

and removed

labels

2026-04-09 07:17:46 +00:00

HAL9000 commented

2026-04-09 07:18:21 +00:00

Author

Owner

Issue triaged by project owner:

State: Verified
Priority: High — Local model support (Ollama/local-llm) is missing from the provider registry. This is a significant gap for users who want to run models locally without API keys.
Milestone: v3.2.0 (already assigned — correct)
Story Points: 5 — L — Adding a new provider type to the registry with proper configuration and testing, 1-2 days.
MoSCoW: MoSCoW/Should have — Local model support is important for developer experience and offline use cases. The spec likely mentions provider extensibility. Should be included in v3.2.0 if capacity allows.
Parent Epic: Needs linking to appropriate v3.2.0 provider Epic

Valid UAT bug. Local model support is a meaningful capability gap.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner

Issue triaged by project owner: - **State**: Verified - **Priority**: High — Local model support (Ollama/local-llm) is missing from the provider registry. This is a significant gap for users who want to run models locally without API keys. - **Milestone**: v3.2.0 (already assigned — correct) - **Story Points**: 5 — L — Adding a new provider type to the registry with proper configuration and testing, 1-2 days. - **MoSCoW**: MoSCoW/Should have — Local model support is important for developer experience and offline use cases. The spec likely mentions provider extensibility. Should be included in v3.2.0 if capacity allows. - **Parent Epic**: Needs linking to appropriate v3.2.0 provider Epic Valid UAT bug. Local model support is a meaningful capability gap. --- **Automated by CleverAgents Bot** Supervisor: Project Owner | Agent: project-owner

HAL9000 referenced this issue

2026-04-09 07:18:38 +00:00

[AUTO-PROJ-OWN] Project Owner Report (Cycle 26) #5296

HAL9000 removed the

MoSCoW

Must have

label

2026-04-09 07:18:51 +00:00

HAL9000 referenced this issue

2026-04-09 07:29:13 +00:00

[AUTO-PROJ-OWN] Project Owner Report (Cycle 1) #5498