Cloudprice MCP

What Albaker-Group/cloudprice-mcp MCP server does

Albaker-Group/cloudprice-mcp MCP server gives AI clients structured cloud pricing and analysis functions across AWS, Azure, GCP, and OCI. Its catalog covers compute, block storage, object storage, managed PostgreSQL, and egress-related comparisons. The repository describes 25 tools in total, including the single-provider price lookups and broader FinOps decision functions.

The server is suited to questions that require comparable monthly or multi-year estimates rather than a single catalog lookup. For example, an agent can compare matching machine sizes across providers, total a workload inventory, estimate migration payback after exit egress, or evaluate commitment options.

How it works

The tools accept structured workload inputs such as vCPUs, memory, quantities, operating hours, storage capacity, disk type, object-storage tier, database size, and egress volume. Results can include hourly and monthly prices, per-cloud totals, sorted alternatives, savings percentages, upfront costs, payback periods, and three-year totals.

Compute lookups target us-east-1 for AWS, eastus for Azure, and us-east1 for GCP. Cross-cloud comparison functions also include OCI. Object storage supports hot, cool, and archive tiers, while managed PostgreSQL pricing separates database compute from storage. OCI Always Free treatment is surfaced for applicable object storage, including up to 20 GB of hot-tier capacity.

The FinOps tools build on these pricing inputs. assess_migration includes one-time exit egress and ranked three-year recommendations; optimize_commitment compares six commitment scenarios; compare_total_cost_of_ownership applies linear year-over-year growth assumptions and provides sensitivity analysis; and find_egress_arbitrage focuses on data-transfer cost differences.

Setup and configuration

Albaker-Group/cloudprice-mcp MCP server requires Python 3.10 or newer. Install the package with pip install cloudprice-mcp, then run cloudprice-mcp setup to detect supported clients and offer configuration changes. The setup command can also target a named client, apply changes without prompting, show a dry run, print configuration, or list detected clients.

After configuration, fully restart the MCP client. The repository also provides cloudprice-mcp doctor for checking the Python version, installation path, configuration location, tool registration, and command path. Manual configuration is available for Windows, macOS, and Linux.

Tools and capabilities

• Look up AWS EC2, Azure VM, and GCP Compute Engine prices.
• Find the least expensive matching compute SKU across four clouds.
• Compare compute inventories, block-storage inventories, and combined workloads.
• Compare object storage and managed PostgreSQL across four providers.
• Apply one-year or three-year compute discount overlays to workload comparisons.
• Model migration cost, commitment options, three-year TCO, and egress arbitrage.

Limitations and notes

Albaker-Group/cloudprice-mcp MCP server returns pricing estimates based on the regions, inputs, and assumptions supported by each tool. Regional scope is explicit for the individual AWS, Azure, and GCP compute lookups. Commitment discounts in compare_workload are fixed overlays of 30% for one year and 50% for three years, while the commitment analysis tool exposes its own six scenarios.

The results should be checked against the requirements of a real deployment. For example, the migration analysis can flag that OCI A1.Flex is ARM and that AMI compatibility should be verified. Pricing comparisons do not, by themselves, establish workload compatibility or account for every provider-specific service condition.