LocalLama MCP Server
by Heratiki
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| LOG_LEVEL | No | Logging level | debug |
| COST_THRESHOLD | No | USD cost above which local offload is preferred | 0.02 |
| OLLAMA_TIMEOUT | No | Ollama timeout in seconds | 120 |
| OLLAMA_ENDPOINT | No | Ollama API base URL | |
| TOKEN_THRESHOLD | No | Token count above which local offload is considered | 1500 |
| QUALITY_THRESHOLD | No | Quality score below which paid API is always used | 0.7 |
| BENCHMARK_PARALLEL | No | Run benchmarks in parallel | false |
| LLAMA_CPP_ENDPOINT | No | llama-server URL; leave unset to disable provider | |
| LM_STUDIO_ENDPOINT | No | LM Studio API base URL | |
| LOCALLAMA_ROOT_DIR | No | Root directory for the server (overrides default project root) | |
| OPENROUTER_API_KEY | No | OpenRouter API key (enables OpenRouter provider) | |
| CODE_SEARCH_ENABLED | No | Enable code search | true |
| DEFAULT_LOCAL_MODEL | No | Model name used when offloading to local provider | |
| MIN_VALIDATOR_SCORE | No | Minimum validation score required before a model is eligible for external validation | 0.6 |
| PROVIDER_TIMEOUT_MS | No | Provider timeout in milliseconds | 120000 |
| OPENROUTER_FREE_ONLY | No | Restrict OpenRouter to free-tier models only | false |
| BENCHMARK_RESULTS_PATH | No | Path to save benchmark results | ./benchmark-results |
| BENCHMARK_SAVE_RESULTS | No | Save benchmark results | true |
| BENCHMARK_TASK_TIMEOUT | No | Benchmark task timeout in milliseconds | 60000 |
| BENCHMARK_RUNS_PER_TASK | No | Number of benchmark runs per task | 3 |
| REMOVE_STALE_LOCK_FILES | No | Automatically remove stale lock files | true |
| VALIDATION_RETRY_BUDGET | No | Validation retry attempts allowed after an initial failed validation | 1 |
| RELIABLE_BENCHMARK_COUNT | No | Benchmark runs required before empirical scores are treated as fully reliable | 3 |
| CODE_SEARCH_INDEX_ON_START | No | Index code on startup | true |
| EXPECT_LOCAL_PROVIDER_DOWN | No | Set true in test to assert no local suggestion | |
| BENCHMARK_MAX_PARALLEL_TASKS | No | Maximum parallel benchmark tasks | 2 |
| CODE_SEARCH_EXCLUDE_PATTERNS | No | JSON array of glob patterns to exclude from indexing | ["node_modules/**","dist/**",".git/**"] |
| CODE_SEARCH_REINDEX_INTERVAL | No | Reindex interval in seconds | 3600 |
| PROVIDER_MAX_CONCURRENT_LOCAL | No | Shared local execution slot count | 1 |
| LOCK_FILE_CHECK_ACTIVE_PROCESS | No | Check for active process when lock file exists | true |
| PROVIDER_MAX_CONCURRENT_REMOTE | No | Per-remote-provider slot count | 5 |
| PROVIDER_HEALTH_PROBE_INTERVAL_MS | No | Provider health probe interval in milliseconds | 60000 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Server capabilities have not been inspected yet.
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
No tools | |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Heratiki/locallama-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server