Allure TestOps MCP server: 13 tools in beta
A server that arrives with the product rather than as a package, what its thirteen tools reach, and the pricing note in its own documentation.
Last verified 27 August 2026 · from Qameta's own Allure TestOps documentation
This page is one of 90 in a directory of monitoring and developer tools. Each page checks what an AI coding agent can get out of the tool through its MCP server, the connector that lets the agent query the tool directly.
Summary
Allure TestOps ships an MCP server inside the product from release 26.1.1. It is exposed at /api/mcp on your own instance, and its documentation lists 13 tools. There is no separate package to install.
The surface is test management rather than test execution: creating and updating test cases, searching them, working with shared steps, reading results and muting.
Two limits come from Qameta's own page. It is a public beta feature, and the documentation states it may be subject to additional fees in future releases.
What is the Allure TestOps MCP server?
It is Allure TestOps's server for the Model Context Protocol. TestOps is a test management platform, so a connected agent can author and search test cases rather than only read results.
The tools split cleanly. testops_create_testcase writes a case with full metadata including steps, tags, custom fields and links, and testops_update_testcase supports partial updates.
Shared steps are first-class. testops_find_sharedsteps, testops_create_sharedstep and testops_update_sharedstep exist because a shared step is reused across many cases, so changing one has wide reach.
Allure TestOps builds its 13-tool server into licensed instances from release 26.1.1. The tools cover test management. They do not expose operational, agent or cost data.
| Part of the work | What Allure TestOps has here | Can an agent reach it |
|---|---|---|
| Code, tests & review | Test cases, shared steps, results, mutes, issues | Yes. Core strength, and the only stage this connector reaches |
| CI & release | Test results uploaded from CI | Partial. testops_find_testresults reads what CI uploaded; nothing triggers or configures a pipeline |
| Production observability | None | No metrics, logs or traces from running services |
| Agent observability & evals | None | No traces or evaluations for your own agents |
| AI cost management | None | No spend data of any kind |
For an agent, that means TestOps answers what the test suite says the product should do, and lets the agent write more of that suite. That is authoring rather than debugging. It has no view of the pipeline or the running system, so both need their own connector.
What can an AI agent do with Allure TestOps?
| Task | Tools | Works | Watch out for |
|---|---|---|---|
| Find an existing test case | testops_find_testcases testops_get_project | Yes | Search is scoped to a project, so establish that first |
| Write a new test case | testops_create_testcase | Yes | Takes full metadata: steps, tags, custom fields and links |
| Amend an existing case | testops_update_testcase | Yes | Partial updates are supported, so a small change stays small |
| Reuse a shared step | testops_find_sharedsteps testops_create_sharedstep testops_update_sharedstep | Yes | A shared step is reused across cases, so an update reaches all of them |
| Read what a run produced | testops_find_testresults | Yes | Results come from what CI uploaded to TestOps |
| Mute a known failure | testops_create_mute testops_delete_mute | Yes | Muting hides a failure from reporting until it is deleted |
| Remove or restore a case | testops_delete_testcase testops_restore_testcase | Yes | Deletion is reversible through the restore tool, which is why both exist |
The agent calls testops_get_project first to establish which project it is in, then composes the search. Almost everything else is scoped to a project, and the wrong project returns an empty result rather than an error.
13 tools documented, shipped with the product
- Orientation2
testops_get_projecttestops_get_issue_details- Finding3
testops_find_testcasestestops_find_sharedstepstestops_find_testresults- Authoring test cases2Write
testops_create_testcasetestops_update_testcase- Authoring shared steps2Write
testops_create_sharedsteptestops_update_sharedstep- Lifecycle2Write
testops_delete_testcasetestops_restore_testcase- Muting2Write
testops_create_mutetestops_delete_mute
Counts come from Qameta's own documentation, read 24 August 2026. The endpoint lives inside a licensed TestOps instance, so there is no package to run and no unauthenticated tools/list to capture.
What connecting costs before the first question
A context window is the amount of text a model can hold at once. Thirteen tools cost 1,104 to 1,138 tokens on Claude. The two tokenizer generations differ by barely thirty tokens, because the descriptions are short.
The server is built into a licensed TestOps instance rather than shipped as a package, so this is what the endpoint on your own instance costs an agent to load.
| Tool set | Tools | Tokens (GPT-5.6, GPT-5.5) | Tokens (GPT-4, GPT-3.5) | Tokens (Claude Haiku 4.5 to Opus 5) | Share of a 200,000-token window |
|---|---|---|---|---|---|
| Every tool documented | 13 | 281 to 408 | 282 to 409 | 1,104 to 1,138 | 0.1% to 0.6% |
Encodings: GPT-5.6, GPT-5.5 o200k_base; GPT-4, GPT-3.5 cl100k_base; Claude Haiku 4.5 to Opus 5 count_tokens. Counted from the 13 tool entries in the vendor's own documentation. That documentation is the only public inventory.
The server could not be run for a live tools/list. No input schemas are available, so these figures cover each tool's name and description only. They are a floor, not the figure.
The Claude columns are measured through the API's count_tokens endpoint with an empty schema per tool. So they cover the same content as the GPT columns. The range there is two tokenizer generations. Haiku 4.5 and Opus 4.6 are at the low end. Opus 5 and Sonnet 5 are at the high end. The gap between them is a small fraction.
Allure TestOps MCP server limits: where answers come back incomplete
The limits below follow from a server that is part of a licensed product rather than a package.
The endpoint does not exist before TestOps 26.1.1.
The documentation states the MCP server is built in and available starting from release 26.1.1, and describes it as a public beta feature.
Guard: Confirm the instance version before planning around this connector, because there is no package to install as a fallback.
The agent says: “I updated one step.” Shared steps are shared.
testops_update_sharedstep changes a step that is reused across test cases, so a single update reaches every case including it.
Guard: Have the agent say how many cases use a shared step before changing it.
The agent says: “The suite is green now.” A mute may be doing that.
testops_create_mute suppresses a failing test from reporting, and testops_delete_mute is what reverses it. Muting is one of the thirteen documented tools.
Guard: Require the agent to report any mute it created alongside a claim that results improved.
How to configure Allure TestOps MCP for agents
- 1
Check your TestOps version
The server exists from release 26.1.1 onward and is not available before it.
- 2
Read the pricing note
The documentation calls this a public beta and says it may be subject to additional fees in future releases.
- 3
Decide who may delete
testops_delete_testcaseis reachable, andtestops_restore_testcaseexists because deletion is recoverable. - 4
Treat shared steps carefully
Updating one changes every case that includes it.
Paste this into your agent’s instructions
Most of this surface writes, and muting hides failures rather than fixing them. An answer must disclose both actions.
When you answer using Allure TestOps MCP tools, state: - The project the answer covers. - Whether you created, updated or deleted a test case or a shared step. A shared step change reaches every case using it. - Whether you muted anything. A mute hides a failure from reporting until it is deleted. Never state that data does not exist. State that no test cases matched the search above. Write your answer in ASD-STE100 Simplified Technical English. Use short sentences with one idea in each.
Do you need the MCP server at all?
Not for a scheduled report. TestOps has a REST API, and for bulk import or anything on a timer it is the direct route.
The connector earns its place during authoring. An agent that has just written code can create the matching test case with steps, tags and links in the same conversation.
For test results across runs rather than test definitions, ReportPortal covers launches and failure history in more depth.
Connect it when an agent is authoring tests. Use the REST API for bulk and scheduled work.
Allure TestOps MCP server setup
There is nothing to install. If your instance runs 26.1.1 or later, the endpoint is already there at /api/mcp.
Clients connect to your own TestOps host, and the instance's own authentication applies.
Claude Code
The token is a TestOps API token from your own instance. Qameta documents one shape for every client: npx mcp-remote bridging to the endpoint over stdio, with a literal api-token scheme rather than Bearer.
claude mcp add allure -- npx mcp-remote \ https://testops.internal/api/mcp \ --header "Authorization: api-token <your-testops-token>"
Cursor
In ~/.cursor/mcp.json. The host is your instance, not a vendor endpoint.
{"mcpServers":{"allure":{"command":"npx","args":[
"mcp-remote","https://testops.internal/api/mcp",
"--header","Authorization: api-token ${AUTH_TOKEN}"],
"env":{"AUTH_TOKEN":"<your-testops-token>"}}}}Codex CLI
Codex reads TOML, so the JSON blocks above will not transfer. The token still applies.
[mcp_servers.allure] command = "npx" args = ["mcp-remote", "https://testops.internal/api/mcp", "--header", "Authorization: api-token <your-testops-token>"]
Every other client
Each block below is the configuration for one client, with the file path and the key that client expects.
Claude Desktop
macOS ~/Library/Application Support/Claude/claude_desktop_config.json. Windows %APPDATA%\Claude\claude_desktop_config.json. There is no CLI. Edit through Settings, Developer, Edit Config.
Quit and restart fully for changes to load. Paths in args must be absolute. Logs at ~/Library/Logs/Claude/mcp.log.
{"mcpServers":{"allure":{"command":"npx","args":[
"mcp-remote","https://testops.internal/api/mcp",
"--header","Authorization: api-token ${AUTH_TOKEN}"],
"env":{"AUTH_TOKEN":"<your-testops-token>"}}}}Gemini CLI
~/.gemini/settings.json globally, or .gemini/settings.json per project. Key mcpServers. CLI: gemini mcp add.
{"mcpServers":{"allure":{"command":"npx","args":[
"mcp-remote","https://testops.internal/api/mcp",
"--header","Authorization: api-token ${AUTH_TOKEN}"],
"env":{"AUTH_TOKEN":"<your-testops-token>"}}}}VS Code Copilot
.vscode/mcp.json per workspace, or your user profile. CLI: code --add-mcp.
{"servers":{"allure":{"command":"npx","args":[
"mcp-remote","https://testops.internal/api/mcp",
"--header","Authorization: api-token ${AUTH_TOKEN}"],
"env":{"AUTH_TOKEN":"<your-testops-token>"}}}}The key is `servers`, not `mcpServers`. VS Code is the only client that uses that name, and copying a config from anywhere else fails silently.
Windsurf
~/.codeium/windsurf/mcp_config.json, which is the path Windsurf documents. Key mcpServers, the same JSON shape as Cursor, carrying the same credential.
Zed
~/.config/zed/settings.json, carrying the same credential as the blocks above.
The key is `context_servers`. Zed does not call them MCP servers in config, so searching its settings for mcp finds nothing.
Google Antigravity
~/.gemini/config/mcp_config.json globally, or .agents/mcp_config.json per project. Key mcpServers. There is no CLI. Use the /mcp overlay in the editor. Supports stdio, streamable HTTP, SSE and websocket.
{"mcpServers":{"allure":{"command":"npx","args":[
"mcp-remote","https://testops.internal/api/mcp",
"--header","Authorization: api-token ${AUTH_TOKEN}"],
"env":{"AUTH_TOKEN":"<your-testops-token>"}}}}Amp (Sourcegraph)
~/.config/amp/settings.json or .amp/settings.json, carrying the same credential. The amp mcp CLI covers approve, doctor and oauth.
The key is `amp.mcpServers`, namespaced. A bare mcpServers block is ignored.
Cline
~/.cline/mcp.json per the docs. The source also reads ~/.cline/data/settings/cline_mcp_settings.json. Key mcpServers, standard JSON shape, carrying the same credential. Supports stdio, streamable HTTP and SSE.
Goose (Block)
~/.config/goose/config.yaml, carrying the same credential.
YAML, and servers are called `extensions`. Goose does not use the MCP vocabulary in config at all. Add one interactively with goose configure, or per session with goose session --with-extension.
Kiro (AWS)
.kiro/settings/mcp.json per workspace, or ~/.kiro/settings/mcp.json globally. Key mcpServers, standard JSON shape, carrying the same credential.
Warp
~/.warp/.mcp.json or .warp/.mcp.json. Key mcpServers, standard shape, carrying the same credential. Also addable through the /agent-add-mcp skill.
JetBrains Junie
.junie/mcp/mcp.json per project, or ~/.junie/mcp/mcp.json. Key mcpServers, carrying the same credential. Use /mcp in the CLI to manage.
Roo Code
.roo/mcp.json per project. The global file is mcp_settings.json, opened from the Roo Code MCP settings view with Edit Global MCP. Key mcpServers, carrying the same credential.
Continue
.continue/mcpServers/*.yaml, one file per server, carrying the same credential.
YAML, and `mcpServers` is a list, not an object. Every JSON client keys servers by name. Continue takes an array, so a converted config will not parse.
Trae
.trae/mcp.json per project, or paste into the UI under Raw Config (JSON). Key mcpServers, standard shape, carrying the same credential.
Devin
Devin has no config file to edit. Servers are added through a web form in the settings UI, and Devin's documentation states you do not need to write or paste JSON. The block below is the shape those fields describe, shown for reference. Either route uses the same credential.
Confirming the version
The feature does not exist before 26.1.1, so this is the first thing to check. The token is unchanged.
# The MCP endpoint is present from TestOps 26.1.1.
curl -H "Authorization: api-token <your-testops-token>" \
https://testops.internal/api/mcpRead-only access, permissions and security
Eight of the thirteen tools write
Creating, updating and deleting test cases and shared steps, plus creating and deleting mutes.
A shared step update has wide reach
Shared steps exist to be reused, so one change propagates to every case that includes it.
Muting hides failures
testops_create_mutesuppresses a failing test from reporting until the mute is removed.The instance is the boundary
The server is part of TestOps, so its authentication and permissions are the ones you already run.
Troubleshooting
- There is no MCP endpoint
- The server ships from release 26.1.1. An earlier TestOps instance does not have it at all.
- A search returns nothing
- Most tools are scoped to a project. Call
testops_get_projectfirst and confirm which project the agent is querying. - A deleted case needs to come back
testops_restore_testcasereverses a deletion, which is why it sits alongside the delete tool.- Failures stopped appearing in reports
- Check for a mute.
testops_create_mutehides a failing test untiltestops_delete_muteremoves the mute.
Allure TestOps MCP server: Reference
| Item | Value |
|---|---|
| Availability | Built into Allure TestOps from release 26.1.1 |
| Status | Public beta, per Qameta's own documentation |
| Pricing note | May be subject to additional fees in future releases |
| Endpoint | /api/mcp on your own TestOps instance |
| Auth | The TestOps instance's own authentication |
| Tools | 13 documented |
| Writing tools | 8 of 13 |
| Inventory source | Qameta's own documentation, read 24 August 2026 |
What engineers report
The documentation is short and specific, so the figures below come from it directly.
| What was checked | What it shows |
|---|---|
| Documented tools | 13 |
| Writing tools | 8 |
| First release | 26.1.1 |
| Status | Public beta |
| Endpoint | /api/mcp on your own instance |
Tool inventory, release version and beta status read from Qameta's own Allure TestOps documentation on 24 August 2026. The endpoint is part of a licensed instance rather than a downloadable package, so no tools/list capture was possible.
Should you connect an agent to Allure TestOps?
Yes. Best for authoring work on a project-scoped token, with mutes treated as a reviewable action.
- Strongest fit
- TestOps teams on 26.1.1 or later who want an agent to author and maintain test cases.
- Main advantage
- Nothing to deploy: the endpoint arrives with the product and uses the authentication you already run.
- Main weakness
- A public beta with a documented possibility of future fees, and no route at all before 26.1.1.
- Operational risk
- Eight writing tools, including deletion and muting, both of which change what reporting shows.
From Oodle
One platform for agent traces and infrastructure
Agent traces usually sit in a different product from the rest of your telemetry, so when a slow database makes an agent slow the symptom is in one tool and the cause is in another. Oodle keeps both in one query surface, from $10 per million spans.
See agent observabilityFrequently asked questions
How many tools does the Allure TestOps MCP server have?
13, read from Qameta's own documentation on 24 August 2026. The endpoint is part of a licensed instance, so there is nothing to capture standalone.
Do I install anything?
No. The server is built into Allure TestOps from release 26.1.1 and served at /api/mcp on your own instance.
Does it cost extra?
It is a public beta. Qameta's documentation states the feature may be subject to additional fees in future releases.
Can an agent delete a test case?
Yes, and restore it. testops_delete_testcase and testops_restore_testcase are both documented tools.
What does muting do?
testops_create_mute hides a failing test from reporting until testops_delete_mute removes the mute.