Provides tools for migrating LLM inference workloads to Arm64 architecture, including workload analysis, migration plan generation, benchmark triggering on Arm GitHub runners, and performance reporting.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@arm-migrate-mcp Analyze this Dockerfile for Arm migration and generate a plan."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Migrate your LLM workload to Arm — and prove the speedup with real numbers.
Arm-Migrate is an MCP server plus a GitHub Actions benchmark harness. Connect it to any MCP client (Claude Code, Claude Desktop, …) and ask it to migrate an LLM inference workload to Arm64. It will:
Analyze the workload (analyze_workload) — scans Dockerfiles,
compose files, and dependency lists for migration blockers: x86-pinned
base images, CUDA-only stacks, missing quantization, arch-specific wheels.
Plan the migration (generate_migration_plan) — emits ready-to-commit
artifacts: an arm64 Dockerfile built around llama.cpp with
KleidiAI kernels, a CI
benchmark workflow, recommended quantization (Q4_0 for KleidiAI's
int8mm/dotprod paths), and a migration checklist.
Measure (trigger_benchmark / fetch_benchmark_results) — runs a
3-way benchmark matrix on GitHub's free ubuntu-24.04-arm runners:
Arm64 + KleidiAI vs Arm64 baseline vs x86 baseline, using llama-bench
with repetitions and captured CPU feature flags.
Report (generate_report) — turns the raw JSON artifacts into a
migration report: prompt-processing and token-generation tokens/sec,
deltas, hardware context, and a go/no-go recommendation.
Zero-cost, fully reproducible: everything runs on free public-repo CI. No GPUs, no cloud account, no API keys.
Qwen2.5-0.5B-Instruct, llama-bench, 5 repetitions, 4 threads. Full reports with stddev and hardware context: Q4_0 · Q8_0.
Comparison | Prompt proc. | Generation |
Arm optimized kernels vs naive Arm build (Q4_0) | +131% | +44% |
KleidiAI vs default kernels (Q4_0) | ~0% | ~0% |
KleidiAI vs default kernels (Q8_0) | +59% | +15% |
Arm64+KleidiAI vs x86 runner (Q8_0) | +269% | +155% |
The practical guidance that falls out: Q4_0 is fast on Arm out of the box
(mainline repack kernels); Q8_0 leaves large gains on the table unless you
build with -DGGML_CPU_KLEIDIAI=ON. Arm64 wins token generation — the
axis that dominates chat/agent serving cost — across every silicon draw we
measured; the x86 prompt-processing picture depends on whether GitHub's
mixed pool hands you AVX-512 (see the variance disclosure in the reports).
Related MCP server: ArmBench MCP Server
Arm64 cloud (Graviton, Axion, Cobalt, Ampere) is routinely the cheapest compute per vCPU, and KleidiAI makes CPU-only LLM inference genuinely usable — but teams don't migrate because they can't predict what their workload gains. Arm-Migrate closes that gap: the agent hands you the migration plan and the measured numbers in one conversation.
npm install && npm run buildRegister with your MCP client (Claude Code shown):
claude mcp add arm-migrate -- node <path>/dist/index.jsThen ask: "Analyze this Dockerfile for Arm migration and generate a plan."
To reproduce our benchmark numbers: fork, enable Actions, run the
arm-bench workflow (workflow_dispatch) — the report lands in the run's
artifacts and job summary.
src/ MCP server (TypeScript)
bench/ benchmark scripts + report generator (no deps)
templates/ generated migration artifacts (Dockerfile.arm64, …)
.github/ the arm-bench harness itselfBuilt for Arm Create: AI Optimization Challenge 2026 (Track 2 — Cloud AI). See RULES.md for the rules digest and DEVLOG.md for an honest build log.
MIT
Analyze an LLM workload (Dockerfile, compose file, dependency list, or plain description) for Arm64 migration blockers and optimization opportunities. Returns severity-ranked signals with concrete fixes.
| Name | Required | Description | Default |
|---|---|---|---|
| compose | No | docker-compose.yml contents | |
| packages | No | requirements.txt / package.json contents | |
| dockerfile | No | Dockerfile contents | |
| description | No | Plain-text description of the workload |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It discloses the output format (severity-ranked signals with fixes) and implies read-only analysis without side effects. However, it does not detail potential rate limits or authentication requirements, but the core behavior is well covered.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, densely packed sentence that front-loads the purpose and scope. Every word adds value, with no redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the 4 optional parameters and no output schema, the description adequately explains the tool's function and output. It lacks details on handling multiple inputs or response format specifics, but is sufficient for an experienced user.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with descriptive parameter names and descriptions. The description adds no additional meaning beyond the schema, so baseline 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool analyzes LLM workloads for Arm64 migration blockers and optimization, specifying input types (Dockerfile, compose, packages, description). It distinguishes from siblings like fetch_benchmark_results or generate_migration_plan which serve different purposes.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies use when analyzing workloads but does not provide explicit guidance on when to use this tool versus alternatives, nor does it mention when not to use it. No comparison with sibling tools is given.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Download artifacts from the latest (or a specific) completed arm-bench run and return the migration report with measured tokens/sec deltas.
| Name | Required | Description | Default |
|---|---|---|---|
| repo | Yes | GitHub repo as owner/name | |
| run_id | No | Specific run ID; defaults to latest completed |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided; description only mentions download and return. Does not disclose authentication needs, rate limits, error handling (e.g., if run incomplete or missing), or whether artifacts are deleted.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Single sentence with no filler. Front-loaded with key action and resource.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
No output schema; description mentions 'migration report with tokens/sec deltas' but not format or structure. Lacks details on error conditions or prerequisites.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema covers both parameters (100% coverage). Description adds 'completed' qualification for run_id and restates schema info, but adds no meaningful extra beyond schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Explicitly states verb 'Download', resource 'artifacts from completed arm-bench run', and output 'migration report with tokens/sec deltas'. Distinct from sibling tools that analyze workloads or plan migrations.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Implicitly suggests use after a completed run, but no explicit 'when to use' or 'when not to use' guidance. No mention of alternatives like analyzing workload instead.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Generate ready-to-commit Arm64 migration artifacts: an arm64 Dockerfile (llama.cpp + KleidiAI), a CI benchmark workflow for free GitHub Arm64 runners, build flags, and a migration checklist.
| Name | Required | Description | Default |
|---|---|---|---|
| quant | No | Target quantization (default Q4_0, the KleidiAI fast path) | |
| model_hint | No | GGUF model URL or HF repo to bake into templates | |
| workload_kind | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It states artifacts are 'ready-to-commit' but does not disclose side effects, permissions, or whether it modifies state.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Single sentence efficiently conveys the tool's purpose and outputs. Could be slightly clearer by breaking into separate points.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
No output schema exists, yet description does not explain return format or how artifacts are provided. Missing guidance on optional parameters.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema covers 67% of parameters with descriptions. The description adds no extra semantic beyond the schema for the parameters.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: generating Arm64 migration artifacts including Dockerfile, CI workflow, build flags, and checklist. It distinguishes from siblings which focus on analysis or benchmarking.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies when to use (for Arm64 migration planning) but does not explicitly contrast with sibling tools or state prerequisites or when not to use it.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Trigger the arm-bench workflow (Arm64+KleidiAI vs Arm64 vs x86 benchmark matrix) on a GitHub repo via workflow_dispatch. Requires an authenticated gh CLI; falls back to manual instructions.
| Name | Required | Description | Default |
|---|---|---|---|
| repo | Yes | GitHub repo as owner/name, e.g. CisnerosCodes/arm-migrate-mcp | |
| model_url | No | Override GGUF model URL for this run |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden. It discloses the requirement for an authenticated gh CLI and the fallback to manual instructions, but omits details about what happens after dispatch (e.g., async behavior, response, error handling). Some behavioral insight is provided, but not comprehensive.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, well-structured sentence that immediately conveys the tool's primary function. Every word adds value; there is no fluff or repetition.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema, the description is not required to explain return values. However, for a workflow trigger tool, it lacks information about expected outcomes (e.g., job URL, status) and how it fits into the pipeline with sibling tools (e.g., use after trigger). It is adequate but not fully complete for an AI agent's decision-making.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 100% coverage: both parameters have descriptions. The tool's description adds context about the workflow (e.g., benchmark matrix) but does not enhance understanding of the parameters beyond what the schema already provides. Baseline score 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action (trigger), the specific workflow (arm-bench with benchmark matrix), and the target (GitHub repo via workflow_dispatch). It distinguishes from sibling tools by focusing on triggering a workflow, which is unique among them.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description mentions a prerequisite (authenticated gh CLI) and a fallback (manual instructions), but does not explicitly state when to use this tool versus its siblings. The usage context is implied by the tool's purpose, but no exclusion criteria or alternative selection guidance is provided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.1.0analyze_workloadfetch_benchmark_resultsgenerate_migration_plantrigger_benchmarkEach tool has a clearly distinct purpose: analyze workloads, fetch results, generate migration plans, and trigger benchmarks. No overlap or ambiguity.
All tool names follow a consistent verb_noun pattern (analyze_workload, fetch_benchmark_results, generate_migration_plan, trigger_benchmark).
With 4 tools, the set covers the core migration workflow without being over- or under-scoped. Each tool earns its place.
The tool surface covers the full migration lifecycle: analyze current workload, benchmark performance, generate migration plan, and trigger benchmarks. No obvious gaps.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
GPU and LLM inference benchmarks, hardware evidence, deployment recommendations, and launch configs.
Measured AI-inference-storage benchmarks with citations, article search, KV-cache ROI estimation.
CPU telemetry & benchmarking, SHA-256 hashing, AES-256-GCM storage, AI video/image via Nova.
Will a local LLM run on your hardware? GGUF quant, buy-vs-rent-vs-API cost, used-GPU prices.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/CisnerosCodes/arm-migrate-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
