Vivado MCP
A Model Context Protocol server that lets AI clients operate AMD Vivado through safe, workflow-level tools while allowing the user to watch and interact with the Vivado GUI.
README
Vivado CLI
Vivado CLI is a CLI-first automation layer for AMD Vivado. It keeps Vivado's
native Tcl as the execution layer, adds persistent session artifacts and
structured JSON output. MCP support has been removed; the product boundary is
now the vivado-cli command and the Vivado Tcl bridge.
The first design target is not GUI click automation. The preferred interactive
mode is a managed Vivado Tcl session that can open the GUI with start_gui,
load a small Tcl bridge, and let vivado-cli submit audited Tcl command files
into that same Vivado process. Batch mode remains useful for CI and fallback
automation.
Current design documents:
- Vivado CLI Design
- ADR 0001: Control Vivado through Tcl batch mode
- ADR 0002: Support managed GUI Tcl sessions
- ADR 0003: Support trusted-local raw Tcl
- ADR 0004: Provide built-in help and skills
- ADR 0005: Remove MCP adapter and standardize on CLI
- Built-in Skills
Initial scope
- Discover a local Vivado installation and report its version.
- Start and stop a managed Vivado Tcl/GUI session.
- Verify whether a requested GUI session has a visible Vivado window without stealing focus, and bring that window forward only when explicitly requested.
- Submit raw Tcl to a managed session when trusted-local expert mode is enabled.
- Create or open project-mode Vivado projects.
- Add RTL/source/constraint files with path validation.
- Audit and manage Vivado filesets (sources / simulation / constraint sets) including include directories, defines, libraries, file properties, top module, USED_IN scopes, dry-run plans, and XDC reorder suggestions.
- Apply structured source-fileset and constraint-set changes with optional before/after state diffs.
- Audit XDC constraint filesets: loading order, per-file command markers, USED_IN scopes, methodology markers, and basic UG903/UG949 sanity warnings.
- Search, dry-run/create, inspect, check upgrade state, upgrade, and generate output products for Vivado project IP.
- Audit simulation setup, dry-run/prepare simulation filesets, launch Vivado simulation, and parse xsim/xelab/xvlog/xvhdl logs into issue IDs.
- Create, inspect, audit, dry-run/mutate, validate, and generate generic IP Integrator block designs.
- Run synthesis, implementation, and bitstream generation.
- Run Non-project Mode flows with audit/dry-run support: read RTL/XDC, check prerequisites, execute synth/opt/place/route, write checkpoints, and collect reports.
- Generate timing, utilization, DRC, methodology, power, and message reports.
- Parse common report outputs into structured summaries and aggregate report diagnostics with issue IDs, root-cause hints, quality gates, next-action plans, and official-document queries.
- Perform explicit hardware access for hw_server targets/devices, debug core/probe discovery, VIO probe readback, generic ILA capture/CSV analysis, and VIO-backed SPI register readback; hardware programming remains out of scope for structured commands.
- Capture JSON state snapshots and diff project/fileset/constraint/IP/BD/run/report state before and after risky or long-running operations.
- Store logs and generated reports as session artifacts.
- Provide built-in help/skills so AI or human CLI callers can learn the intended Vivado workflows before acting.
- Package AMD official Vivado documentation metadata and topic guidance as the authority layer for help and expert Tcl planning.
Capability profiles
safe: workflow tools only; no raw Tcl.trusted-local: workflow tools plus raw Tcl/source-file execution inside the managed Vivado session.unrestricted: raw Tcl with minimal policy checks for personal local use.
The packaged bridge in src/vivado_cli/assets/cli_bridge.tcl is the core control path: vivado-cli submits Tcl files to a live Vivado Tcl/GUI session and receives result files back.
Built-in help
The CLI exposes tutorial and authority content through JSON commands:
vivado-cli help topic <topic>vivado-cli assist next --goal "<task>"vivado-cli skills listvivado-cli skills get <skill_id>vivado-cli tools listvivado-cli tools describe <command-or-tool-id>vivado-cli tcl help <command>vivado-cli tcl review --file <script.tcl>
Seed skill docs live in docs/skills.
The official reference layer stores document IDs, AMD URLs, scope summaries, topic routing, and local filename candidates. It does not copy the full AMD document text into this repository.
Install For Local Use
From a fresh clone:
git clone https://github.com/Shanqianlvshui/Vivado_CLI.git
cd Vivado_CLI
python -m venv .venv
.\.venv\Scripts\python.exe -m pip install -e ".[dev]"
Basic smoke checks after installation:
.\.venv\Scripts\vivado-cli.exe --help
.\.venv\Scripts\vivado-cli.exe tools list
.\.venv\Scripts\vivado-cli.exe tcl help create_clock
This machine has been tested with:
C:\Xilinx\Vivado\2023.1\bin\vivado.bat
The stable local entry point for agents and external tools is:
C:\Tools\vivado-cli\bin\vivado-cli.exe
New terminals can also read the full path from:
$env:VIVADO_CLI_EXE
Fresh Clone Verification
To verify the published repository from scratch, run:
powershell -ExecutionPolicy Bypass -File .\scripts\verify-cold-clone.ps1
The default smoke path clones https://github.com/Shanqianlvshui/Vivado_CLI.git,
installs the editable package in a temporary virtual environment, runs CLI help
and tool-discovery checks, compiles src and tests, and runs the fast
fake-Vivado smoke tests. Add -Full to run the full unit suite in the cold
clone.
CLI Usage
vivado-cli is the primary entry point. It writes persistent session records
under .vivado_cli/sessions, so separate CLI invocations can operate the same
live Vivado process.
vivado-cli check-installation --vivado-path C:\Xilinx\Vivado\2023.1\bin\vivado.bat
vivado-cli --workspace C:\Workspace\Vivado_mcp session start `
--vivado-path C:\Xilinx\Vivado\2023.1\bin\vivado.bat
vivado-cli --workspace C:\Workspace\Vivado_mcp session recovery `
--session <session_ref>
vivado-cli --workspace C:\Workspace\Vivado_mcp session timeline `
--session <session_ref> `
--limit 20
vivado-cli --workspace C:\Workspace\Vivado_mcp session artifacts `
--session <session_ref> `
--kind report `
--limit 10
vivado-cli --workspace C:\Workspace\Vivado_mcp session read-artifact `
--session <session_ref> `
reports\timing_summary.rpt `
--max-chars 12000
vivado-cli --workspace C:\Workspace\Vivado_mcp session open-project `
--session <session_ref> `
C:\Workspace\Vivado\XCZU19EG\XCZU19EG_TEST\projects\qt7331_adda_2023.1\qt7331_adda_2023.1.xpr
vivado-cli --workspace C:\Workspace\Vivado_mcp bd summary `
--session <session_ref> `
--design jesd204b_bd `
--validate
vivado-cli --workspace C:\Workspace\Vivado_mcp run status `
--session <session_ref> `
--run synth_1
vivado-cli --workspace C:\Workspace\Vivado_mcp run launch `
--session <session_ref> `
synth_1 `
--jobs 8
vivado-cli --workspace C:\Workspace\Vivado_mcp run launch-local `
--session <session_ref> `
--jobs 8 `
synth_1
vivado-cli --workspace C:\Workspace\Vivado_mcp run diagnose `
--session <session_ref> `
synth_1
vivado-cli --workspace C:\Workspace\Vivado_mcp run logs `
--session <session_ref> `
synth_1 `
--tail 80
vivado-cli --workspace C:\Workspace\Vivado_mcp run reset `
--session <session_ref> `
synth_1 `
--expect-destructive
vivado-cli --workspace C:\Workspace\Vivado_mcp hw list-debug-cores `
--session <session_ref> `
--expect-hardware-access
vivado-cli --workspace C:\Workspace\Vivado_mcp hw vio-read `
--session <session_ref> `
--vio hw_vio_0 `
--probe chip_config/spi_read_status `
--probe chip_config/spi_read_req `
--expect-hardware-access
vivado-cli --workspace C:\Workspace\Vivado_mcp hw vio-write `
--session <session_ref> `
--vio hw_vio_0 `
--set chip_config/spi_read_req=0 `
--expect-hardware-access `
--expect-vio-write
vivado-cli --workspace C:\Workspace\Vivado_mcp hw capture-ila `
--session <session_ref> `
--ila hw_ila_0 `
--depth 1024 `
--analysis adc14 `
--sample-rate-hz 312500000 `
--label bringup_capture `
--expect-hardware-access
vivado-cli --workspace C:\Workspace\Vivado_mcp hw spi-read `
--session <session_ref> `
--vio hw_vio_0 `
--status-probe spi/status `
--req-probe spi/req `
--target-probe spi/target `
--addr-probe spi/addr `
--reg 2:0x0281 `
--reg 2:0x0300 `
--expect-hardware-access
vivado-cli --workspace C:\Workspace\Vivado_mcp tcl review `
--file .\scripts\change_bd.tcl
vivado-cli --workspace C:\Workspace\Vivado_mcp session run-tcl `
--session <session_ref> `
--file .\scripts\change_bd.tcl `
--expect-destructive
Environment
VIVADO_CLI_WORKSPACE is the default workspace for managed sessions.
VIVADO_CLI_ALLOWED_ROOTS is a semicolon-separated list on Windows; workflow
paths such as projects, sources, constraints, and Tcl files in trusted-local
mode must stay under one of these roots. VIVADO_CLI_DOCS_ROOT points to the
local AMD Vivado documentation library used by the official-reference index; it
defaults to C:\Database\domains\fpga\xilinx\vivado\docs\raw. Set
VIVADO_CLI_PDFTOTEXT if pdftotext is not on PATH.
AI Operating Flow
CLI callers should use this order:
- Start with
vivado-cli assist next --goal "<task>"and include--session,--last-error, or a Tcl draft with--tcl/--filewhen available. - Discover the available surface with
vivado-cli tools list,vivado-cli tools describe <command>,vivado-cli skills list, andvivado-cli skills get <skill_id>. - Use
vivado-cli tcl help <command>before unfamiliar Vivado Tcl; it combines official-document search, CLI coverage guidance, and optional installed Vivado help when a session is attached. - Use structured commands first:
project summary,fileset ...,constraint ...,bd ...,run ...,report, andhw .... - Use
vivado-cli tcl reviewbefore raw expert Tcl, thenvivado-cli session run-tclorvivado-cli session source-tclonly when no structured command covers the task. - Pass explicit acknowledgements for risky actions, such as
--expect-destructive,--expect-hardware-access, and--expect-vio-write. - Read
state_trackingandstate_difffrom mutating fileset/constraint commands before launching long runs or handing the session to another agent. - For a resumed or stale thread, run
vivado-cli session recovery, then inspectsession timeline,session artifacts, orsession read-artifactbefore changing Vivado state. - After mutating project state, refresh with
vivado-cli project summary,vivado-cli fileset describe,vivado-cli constraint check-order,vivado-cli bd summary, orvivado-cli run diagnoseas appropriate.
First Manual Test
Use this sequence:
vivado-cli help topic gui-sessionvivado-cli tools listvivado-cli check-installation --vivado-path C:\Xilinx\Vivado\2023.1\bin\vivado.batvivado-cli session start --vivado-path C:\Xilinx\Vivado\2023.1\bin\vivado.batvivado-cli session state --session <session_ref>vivado-cli tcl help create_projectvivado-cli tcl review --tcl "return \"version=[version -short]\""vivado-cli session run-tcl --session <session_ref> --tcl "return \"version=[version -short]\""vivado-cli project summary --session <session_ref>after opening or creating a projectvivado-cli session stop --session <session_ref>
Implemented CLI Commands
Use vivado-cli tools list as the source of truth. The current top-level
groups are:
check-installationsession start|adopt|list|state|artifacts|timeline|read-artifact|recovery|open-project|run-tcl|source-tcl|stoptcl help|reviewskills list|gethelp topicassist nexttools list|describeproject summaryfileset list|describe|create|add-files|remove-files|set-file-properties|set-top|applyconstraint diagnostics|check-order|applybd summary|validaterun status|launch|launch-local|logs|diagnose|resetreporthw list-debug-cores|vio-read|vio-write|capture-ila|spi-read
Development Checks
.\.venv\Scripts\python.exe -m compileall -q src tests
.\.venv\Scripts\python.exe -m pytest tests\unit -q
The test suite includes a fake Vivado process and CLI lifecycle tests. GitHub
Actions runs the same package install, compile check, CLI smoke, and full unit
suite on windows-latest.
Artifacts
Command files, result files, logs, and reports are stored under the managed session directory and exposed through artifact URIs:
vivado://sessions/{session_ref}/artifacts/{artifact_id}
Use vivado-cli session artifacts --session <session_ref> to list artifact IDs
and vivado-cli session read-artifact --session <session_ref> <artifact_id> to
read a bounded text slice. vivado-cli session recovery --session <session_ref>
returns the latest analyses, snapshots, quality gates, timeline preview, and
next CLI actions for AI handoff and long-running task recovery.
Commands that generate reports, summaries, captures, snapshots, or diffs return
filesystem paths and vivado://... artifact URIs in their JSON output.
Structured fileset and constraint mutations attach state_tracking and
state_diff by default; pass --no-state-diff only for intentional bulk edits
where speed matters more than immediate audit artifacts.
Official Reference Resources
Use vivado-cli tcl help <command> for Tcl command routing. It searches the
local AMD/Xilinx documentation library when available, reports the official-doc
topic, shows any structured CLI coverage for that command family, and can also
query installed Vivado help when --session <session_ref> is supplied.
The CLI uses the local PDF library under
C:\Database\domains\fpga\xilinx\vivado\docs\raw by default. Set
VIVADO_CLI_DOCS_ROOT to another documentation root and VIVADO_CLI_PDFTOTEXT
if pdftotext is not on PATH.
Explicitly out of scope for the first version
- GUI click automation.
- Attaching to an arbitrary already-open Vivado process that did not load the CLI bridge.
- Hardware programming, configuration-memory writes, boot operations, and debug/probe mutation. Read-only hardware discovery is supported with explicit confirmation.
- Advanced IP Integrator automation beyond the generic BD action model.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.