IntegrationUpdated
Connect Codex to your backlog
Stride is a backlog MCP server for Codex. Model Context Protocol (MCP) is an open standard that lets Codex use approved tools and context from Stride, so it can work from linked stories, decisions, diagrams, processes, and tests instead of reconstructing intent from a repository alone.
How to connect Codex
Sign in with your Stride account and there is nothing to copy. For CI and other runs that cannot open a browser, use an agent key instead.
Sign in with your Stride account
- Add Stride to Codex:
codex mcp add stride --url https://www.stride.page/api/mcp/stride. - Sign in:
codex mcp login stride. Your browser opens to Stride. - Sign in, choose the project Codex works in and the projects it can reach, and click Allow. You can disconnect it at any time in Settings → Integrations → Connected apps.
Or install the Stride plugin: $stride:next and $stride:done
The plugin adds the same server and two skills. $stride:next picks up your next story (the one a teammate staged for you, or the top one assigned to you) with its full spec and starts it. $stride:done checks the work against the story’s acceptance criteria and moves it to In Review.
It is listed as Stride AI in OpenAI’s plugin directory, which Codex shares with ChatGPT. Start codex, run /plugins and install Stride AI by Newlight Solutions, or open Plugins in the ChatGPT desktop app. If Codex asks you to connect it, sign in to Stride and allow it. Start a new session and run $stride:next.
Or add the plugin’s own marketplace, from GitHub:
codex plugin marketplace add Newlight-Solutions/stride-pluginsThen start codex, run /plugins, open Stride AI in that marketplace and install it. Codex asks you to sign in to Stride. Start a new session and run $stride:next.
Or use an agent key
- In Stride, open Settings → Integrations → Codex and generate a project-scoped agent key.
- Export it as
STRIDE_AGENT_KEYin the environment that launches Codex. - Add the TOML configuration below to
~/.codex/config.toml, or to.codex/config.tomlin a trusted project.
[mcp_servers.stride]
url = "https://www.stride.page/api/mcp/stride"
bearer_token_env_var = "STRIDE_AGENT_KEY"
tool_timeout_sec = 300Or add it with one command, which writes the same entry to ~/.codex/config.toml: codex mcp add stride --url https://www.stride.page/api/mcp/stride --bearer-token-env-var STRIDE_AGENT_KEY. Then run codex mcp list, or /mcp inside Codex, to check that Stride is connected.
Good to know
- One configuration, three apps. The Codex CLI, its IDE extension and the ChatGPT desktop app share this configuration. Restart whichever you use after changing it; with an agent key, start it from a place where
STRIDE_AGENT_KEYis set. - Project configuration needs a trusted project. Codex reads a project’s
.codex/config.tomlonly once you trust the project. The file holds the endpoint and, with a key, only the variable’s name, so it can be committed and each developer signs in or uses their own key. - No timeout to change. Codex gives each tool call 60 seconds by default. Stride answers within that, and long generations such as a diagram carry on in the background, where Codex can check on them. The configuration above allows 300 seconds anyway, Stride’s own limit.
- It runs where your configuration is. ChatGPT on the web does not read your local Codex configuration, so connect Stride from the CLI, the IDE extension or the desktop app. For ChatGPT itself, install Stride AI from its plugin directory: see the ChatGPT guide.
- A CI runner needs a key.
codex execon a fresh machine cannot open a browser to sign in, so give it an agent key throughbearer_token_env_var.
Checked against OpenAI’s Codex documentation on October 26, 2026.
What Codex can do
Stride exposes 131 unique, externally callable tools. Access is constrained by the member, workspace, project scope, and role behind the sign-in or the agent key.
| Module | Representative work |
|---|---|
| Plan | Read connected story context; create and update stories, epics, sprints, and releases. |
| Design | Read diagrams and architecture decisions; create, update, and link documents and ADRs. |
| Optimize | Read processes, process insights, delivery KPIs, and linked modernization context. |
| Verify | Create test cases, inspect test runs and defects, and run release quality-gate checks. |
Frequently asked questions
How do I connect Codex to Stride?
Run codex mcp add stride --url https://www.stride.page/api/mcp/stride, then codex mcp login stride. You sign in to Stride in your browser, choose the projects Codex can reach, and allow it, so there is no key to copy. Or install the Stride AI plugin from OpenAI’s plugin directory (run /plugins in Codex), which adds the same server plus $stride:next and $stride:done. For CI and other runs that cannot open a browser, generate an agent key in Settings → Integrations → Codex, export it as STRIDE_AGENT_KEY, and run codex mcp add with --bearer-token-env-var STRIDE_AGENT_KEY instead. Run codex mcp list to check it.
What does the Codex configuration contain?
Signed in, it names the Stride MCP endpoint and nothing else: Codex keeps the sign-in itself. With an agent key, it also tells Codex to read its bearer token from STRIDE_AGENT_KEY, so the secret stays in your environment instead of being written into config.toml.
Can Codex change project data?
Yes, as far as your permissions and the projects it was allowed to reach permit. Every tool call is checked by Stride before its handler runs, and writes are recorded as MCP activity for the person who connected it.
Does the same connection work with Claude Code and Cursor?
All three clients use the same Stride MCP endpoint and tool registry, and sign in the same way. Their configuration formats differ, so use the matching setup guide. Each app you sign in is listed on its own under Connected apps.
Which plan includes the MCP server?
Every paid plan includes MCP access, the public API, and webhooks. Starter includes one API or agent key and one webhook; Pro and Enterprise increase those limits. An app you sign in uses one agent key. The pricing page contains the current plan details.
Other integrations
- GitHubSource control · GitHub → Stride
- GitLabSource control · GitLab → Stride
- SlackCommunication · Stride → Slack
- Jira CloudProject management · 30-day two-way
- ClaudeAI assistant · MCP read/write
- ChatGPTAI assistant · MCP read/write
- Claude CodeAI coding agent · MCP read/write
- CursorAI coding agent · MCP read/write
- Webhooks & REST APIBuild your own · REST + webhooks
- How AI writes acceptance criteria (and where it fails)The honest map of where AI is dramatically better than humans at writing acceptance criteria, and the five places it confidently writes garbage. Plus the prompts that work.10 min read
- The connected delivery graph: one source of truth from PRD to prodMost teams ship software with five tools that don't talk to each other. The friction isn't any individual tool. It's the missing graph between them. This is the case for one connected graph.9 min read
- AGENTS.md and CLAUDE.md for coding agentsWhich agents read AGENTS.md or CLAUDE.md, how to keep one file for all of them, what belongs in it, and how to check it loaded when an agent ignores it.11 min read
- Backlog access for coding agents over MCPWhat a backlog MCP server should give a coding agent, how to scope its key, why tickets are untrusted input, and the config mistakes that fail silently.8 min read
- Writing stories a coding agent can executeWhat an agent needs that a teammate would infer: the outcome, the reason, testable criteria, what not to touch, and the check that proves it done.9 min read
- Reviewing pull requests from coding agentsA working method for reviewing agent-written code: criteria before the diff, the failure modes to look for, the evidence to demand, and a second agent.7 min read