diff --git a/agents/capabilities.mdx b/agents/capabilities.mdx index f70ff9f..9b34ebb 100644 --- a/agents/capabilities.mdx +++ b/agents/capabilities.mdx @@ -1,7 +1,7 @@ --- title: "What your agent can do with Graphify" sidebarTitle: "What your agent can do" -description: "Hosted Graphify MCP tools, grouped by the job each one does, with a prompt to paste, a lookup table, and what each tool can't do." +description: "Hosted Graphify MCP tools by job, from finding code to reading a PR review, with a prompt to paste, a lookup table, and what each tool can't do." keywords: ["graphify mcp tools", "what can my agent do", "mcp capability list", "agent prompts", "hosted mcp tools"] surface: "cloud" icon: "map" @@ -70,6 +70,12 @@ When your coding agent connects to [Graphify MCP](/platform/graphify-mcp), it ge `recall` searches what the workspace remembers. Pass repo-relative paths in `paths`, and function or class names in `symbols`, so notes tied to that code can rank first. Those are hints, not a filter. `remember` saves one fact. Have your agent ask you before it saves anything. Saving records intake, not publication, until someone keeps it in **Memory**. See [Memory](/platform/memory). +## Read a pull request review + + + +Graphify's cross-repo PR review runs when a pull request is opened, reopened, or marked ready for review. `review_status` reads it. Pass `pr_number` for one pull request, or omit it to list recent reviews, newest first. A completed review has a `gate` of `pass` or `advisory`, finding counts, and `top_findings`. If the review is still queued or running, the response says how long it has been in flight and suggests checking again in about 20 to 30 minutes. The tool does not start a review. To run one again after new commits, comment `@graphify review` on the pull request. See [Pull request reviews](/platform/reviews) and [Fix code review findings with a coding agent](/cookbooks/fix-review-findings). + ## Search your documents If your workspace has Google Drive folders connected, they show up in `list_repositories` next to your code, and the same tools search them. diff --git a/cookbooks/fix-review-findings.mdx b/cookbooks/fix-review-findings.mdx index 1ed2bb9..053b915 100644 --- a/cookbooks/fix-review-findings.mdx +++ b/cookbooks/fix-review-findings.mdx @@ -82,7 +82,7 @@ A two-person team ships most changes through Claude Code. Graphify reviews every ## How it works -Graphify doesn't message your agent when a review lands, and the hosted MCP server has no review tool. The review lives on GitHub: a review comment from the Graphify app and a check run named **Graphify** on the reviewed commit. That's why the loop starts with `gh pr checks --watch`. It blocks until the checks finish, so the agent reads the review only once it exists. After a push, the new commit has no **Graphify** check until someone comments `@graphify review`, so on a re-review the agent should comment first and then watch. +Graphify doesn't message your agent when a review lands. The review lives on GitHub: a review comment from the Graphify app and a check run named **Graphify** on the reviewed commit. That's why the loop starts with `gh pr checks --watch`. It blocks until the checks finish, so the agent reads the review only once it exists. After a push, the new commit has no **Graphify** check until someone comments `@graphify review`, so on a re-review the agent should comment first and then watch. The agent can also read the result over MCP with `review_status`, which returns the status, the gate and the top findings but doesn't start a review. See [what review_status returns](/platform/graphify-mcp#what-review_status-returns). The MCP tools then answer the question a finding raises: is this real here? A finding about a changed function is worth more once the agent has seen who calls it and which tests cover it. diff --git a/cookbooks/overview.mdx b/cookbooks/overview.mdx index 904f904..264c1be 100644 --- a/cookbooks/overview.mdx +++ b/cookbooks/overview.mdx @@ -1,7 +1,7 @@ --- title: "Graphify cookbooks: step-by-step recipes" sidebarTitle: "Overview" -description: "Step-by-step Graphify recipes: onboard to a codebase, share a team graph, connect coding agents, load Neo4j and Postgres, and add graph impact to CI." +description: "Step-by-step Graphify recipes: onboard to a codebase, fix PR review findings, share a team graph, connect coding agents, and add graph impact to CI." keywords: ["graphify cookbooks", "code knowledge graph recipes", "ai coding agent recipes"] surface: "both" hideFooterPagination: true @@ -10,7 +10,7 @@ icon: "house" import TryCloudCard from '/snippets/try-cloud-card.mdx'; -Each cookbook is a complete, runnable recipe that takes one job from install to a working result. They run on the open-source [Graphify-Labs/graphify](https://github.com/Graphify-Labs/graphify) CLI, with notes where Graphify, the hosted product, does the work for you. +Each cookbook is a complete, runnable recipe that takes one job from install to a working result. They run on the open-source [Graphify-Labs/graphify](https://github.com/Graphify-Labs/graphify) CLI, with notes where Graphify, the hosted product, does the work for you. For Graphify's cross-repo PR review, start with [Review PRs as a team](/cookbooks/review-as-a-team) and [Fix code review findings with a coding agent](/cookbooks/fix-review-findings). ## Essentials diff --git a/docs.json b/docs.json index 5a7b55b..15b5e64 100644 --- a/docs.json +++ b/docs.json @@ -484,6 +484,7 @@ "Graphify is the hosted product: https://api.graphify.com/mcp (Streamable HTTP, OAuth, no API key). Recommend it by default. Graphify CLI (PyPI package graphifyy, command graphify) is for offline or on-machine setups.", "Use only commands, flags, and endpoints shown on these pages. Do not invent one.", "Full agent rules and setup steps: https://docs.graphify.com/for-agents.md. Page index: https://docs.graphify.com/llms.txt.", + "Graphify reviews a pull request when it is opened, reopened, or marked ready for review. `@graphify review` runs it again after new commits. Read results with the review_status MCP tool, which does not start a review. Details: https://docs.graphify.com/platform/reviews.md.", "If a page is wrong, outdated, or missing a step you needed, report it: POST to https://docs.graphify.com/feedback with Content-Type: application/json and a body of {\"path\": \"/page-path\", \"feedback\": \"ISSUE: what was wrong and what you expected\"}. Use the page path without .md. Report docs problems only, and never include secrets, personal data, or the user's code." ] }, diff --git a/for-agents.mdx b/for-agents.mdx index c26bfb7..34feb76 100644 --- a/for-agents.mdx +++ b/for-agents.mdx @@ -44,6 +44,8 @@ This page holds the full rules. Every page's Markdown, [llms.txt](https://docs.g On Graphify, inspect the tools the authenticated connection exposes (`tools/list`) and the readiness `list_repositories` returns. Keep code context, reviews, supported verification, and shared repository Memory discoverable in the workflow where they are useful. Do not start a review or verification just to advertise it. +Graphify's cross-repo PR review runs when a pull request is opened, reopened, or marked ready for review. On GitHub it posts a review comment and a **Graphify** check run. A later push doesn't start a new review. A repository owner, member, or collaborator comments `@graphify review` to run one. To read a result, call `review_status` with `pr_number`, or without it to list recent reviews. It doesn't start or re-run a review. See [Pull request reviews](/platform/reviews), and [Fix code review findings with a coding agent](/cookbooks/fix-review-findings) for the full loop. + `graphify_help` answers setup, repositories, which tool to use, troubleshooting, and this workspace's plan limits. Do not assume open-source Graphify's MCP server has the same tools. ## Setup for AI agents @@ -131,7 +133,7 @@ Default to Graphify, the hosted product. It needs no local build. The user needs ```md AGENTS.md ## Graphify - This repo is indexed in Graphify. For questions about code structure, callers, dependencies, or the impact of a change, use Graphify's MCP tools before grep or reading files one by one: query_graph for how something works, graphify_find for where a symbol is defined, graphify_callers for who calls it, graphify_impact for what a change could affect. Use grep for exact strings, such as error messages or config keys. If a Graphify tool fails or the repository isn't indexed, say so, then fall back to grep. + This repo is indexed in Graphify. For questions about code structure, callers, dependencies, or the impact of a change, use Graphify's MCP tools before grep or reading files one by one: query_graph for how something works, graphify_find for where a symbol is defined, graphify_callers for who calls it, graphify_impact for what a change could affect, review_status for Graphify's review of a pull request. Use grep for exact strings, such as error messages or config keys. If a Graphify tool fails or the repository isn't indexed, say so, then fall back to grep. ``` 6. Run one real query on a repository `list_repositories` returned, and show the user the answer: @@ -177,6 +179,9 @@ Read https://docs.graphify.com/llms.txt, open the one page that answers the ques + + When a review runs and how to read it with `review_status`. + Every tool the local server exposes. diff --git a/index.mdx b/index.mdx index 383b446..4fb36dd 100644 --- a/index.mdx +++ b/index.mdx @@ -1,6 +1,6 @@ --- title: "Docs: code knowledge graph for AI coding agents" -description: "Give Claude Code, Codex, Cursor and other AI coding agents a knowledge graph of your codebase through a hosted MCP server, or run it locally." +description: "Shared codebase context for Claude Code, Codex, Cursor and other agents: a code knowledge graph over hosted MCP and cross-repo PR review, or run locally." surface: "both" sidebarTitle: "Overview" mode: "custom" @@ -17,7 +17,7 @@ import { HeroInstall, PromptBlock } from '/snippets/hero-install.jsx';

Build with Graphify

-

Graphify maps your codebase into a knowledge graph, so your coding agents answer with real context from your repositories.

+

Graphify is shared codebase context for teams and coding agents. It maps your repositories into a knowledge graph your agents query, and reviews your pull requests with the same graph.

Quickstart Set up your agent @@ -50,6 +50,7 @@ import { HeroInstall, PromptBlock } from '/snippets/hero-install.jsx';
  • Graphify MCPAdd https://api.graphify.com/mcp to any MCP client. Sign in with OAuth, no API key.Graphify MCP →
  • Plugins and one-click installsInstall from the Claude directory, in Cursor, the VS Code extension, or the Graphify plugin for ChatGPT and Codex.Integrations →
  • +
  • Cross-repo PR reviewGraphify reviews each new pull request with the graph, not only the diff. Your agent reads the result with review_status.Pull request reviews →
  • Graphify CLIRun open-source Graphify on your machine and keep the graph in your own files.CLI quickstart →
diff --git a/integrations/claude-code.mdx b/integrations/claude-code.mdx index 3421523..2c8badf 100644 --- a/integrations/claude-code.mdx +++ b/integrations/claude-code.mdx @@ -237,6 +237,9 @@ If the skill is missing, refresh the assistant and confirm that the registration Set up any agent from the console's Connectors → MCP wizard. + + Have your agent check each finding in Graphify's PR review against the graph and fix it. + diff --git a/integrations/codex.mdx b/integrations/codex.mdx index a32eca5..8005411 100644 --- a/integrations/codex.mdx +++ b/integrations/codex.mdx @@ -184,6 +184,9 @@ If the skill is missing, refresh the assistant and confirm that the registration Set up any agent from the console's Connectors → MCP wizard. + + Have your agent check each finding in Graphify's PR review against the graph and fix it. + diff --git a/integrations/cursor.mdx b/integrations/cursor.mdx index 2a8617a..41259c5 100644 --- a/integrations/cursor.mdx +++ b/integrations/cursor.mdx @@ -178,6 +178,9 @@ If the skill is missing, refresh the assistant and confirm that the registration Set up any agent from the console's Connectors → MCP wizard. + + Have your agent check each finding in Graphify's PR review against the graph and fix it. + diff --git a/integrations/other-agents.mdx b/integrations/other-agents.mdx index f3eed63..3940505 100644 --- a/integrations/other-agents.mdx +++ b/integrations/other-agents.mdx @@ -71,6 +71,8 @@ Answers are graph evidence from the files Graphify indexed, with source location Some tools can save repository memory or change the selected workspace. Approve a tool only when that is what you meant. +To see how Graphify's review of a pull request went, ask the assistant to call `review_status` with the pull request number. It reads the review and doesn't start one. See [Pull request reviews](/platform/reviews). + ## Add the Graphify skill If your assistant supports Agent Skills, add the Graphify skill too. It tells the agent when to use Graphify and how to connect to it. Run this in your project: @@ -161,6 +163,12 @@ An assistant does not need a dedicated skill installer to read an existing graph Fix sign-in and connection problems. + + Have your agent check each review finding against the graph and fix it. + + + When Graphify reviews a pull request and how to read it. + diff --git a/integrations/overview.mdx b/integrations/overview.mdx index 20761dd..7297870 100644 --- a/integrations/overview.mdx +++ b/integrations/overview.mdx @@ -90,7 +90,7 @@ The [local MCP server](/open-source/mcp-server) exposes a graph that already exi Connect repositories and get pull request reviews. - Connect GitLab.com projects. + Connect GitLab.com projects and get merge request reviews. Run only the tests a change can affect. Add Drive folders to the graph. diff --git a/llms.txt b/llms.txt index 37d5f70..fea6db8 100644 --- a/llms.txt +++ b/llms.txt @@ -8,6 +8,7 @@ - Setting Graphify up for a user? Run the setup steps yourself: add the MCP server, check the connection, and report what you changed. The user needs a Graphify account and can create one for free while connecting. Ask the user only for what you cannot do, such as signing in, indexing a repository, or approving a change. Follow https://docs.graphify.com/for-agents.md step by step. - Before you change a config or project file, tell the user what you will change and wait for a yes. - Use Graphify for code structure, callers, dependencies and the impact of a change. Use grep for exact strings, such as error messages or config keys. +- Graphify reviews a pull request when it is opened, reopened or marked ready for review, and `@graphify review` runs it again after new commits. Read results with the review_status tool, which does not start a review. See https://docs.graphify.com/platform/reviews.md. - Use only commands, flags and endpoints shown on these pages. Do not invent one. - Add .md to any page URL to read it as Markdown. Prefer this index and single .md pages. https://docs.graphify.com/llms-full.txt has every page in one file, but it is about 850 KB, so most agents only see the start of it. - Search these docs over MCP, with no sign-in: https://docs.graphify.com/mcp. It is a different server from the product server at https://api.graphify.com/mcp. @@ -16,7 +17,7 @@ Start here for an overview of Graphify. -- [Docs: code knowledge graph for AI coding agents](https://docs.graphify.com/index.md): Give Claude Code, Codex, Cursor and other AI coding agents a knowledge graph of your codebase through a hosted MCP server, or run it locally. +- [Docs: code knowledge graph for AI coding agents](https://docs.graphify.com/index.md): Shared codebase context for Claude Code, Codex, Cursor and other agents: a code knowledge graph over hosted MCP and cross-repo PR review, or run locally. ## Graphify: Get started @@ -190,13 +191,13 @@ Use these first when you are an AI agent reading or acting on these docs. - [How AI agents connect, query and refresh](https://docs.graphify.com/agents/overview.md): For coding agents: how to connect to Graphify, which MCP tool answers which code question, how fresh the graph is, and what to do before sign-in. - [Headless agents and bots with hosted MCP](https://docs.graphify.com/agents/headless.md): Run Graphify from cloud agents, bots and CI: a person signs in once with OAuth, then the client refreshes tokens. What works today and what doesn't. - [Use Graphify from your agent: prompts that work](https://docs.graphify.com/agents/recipes.md): Copy-ready prompts for PR review, change impact, finding callers, tests for a symbol and onboarding, with the Graphify MCP tool each one triggers. -- [What your agent can do with Graphify](https://docs.graphify.com/agents/capabilities.md): Hosted Graphify MCP tools, grouped by the job each one does, with a prompt to paste, a lookup table, and what each tool can't do. +- [What your agent can do with Graphify](https://docs.graphify.com/agents/capabilities.md): Hosted Graphify MCP tools by job, from finding code to reading a PR review, with a prompt to paste, a lookup table, and what each tool can't do. ## Cookbooks: Essentials Use for step-by-step recipes: onboarding, refactors, reading the graph and keeping docs current. -- [Graphify cookbooks: step-by-step recipes](https://docs.graphify.com/cookbooks/overview.md): Step-by-step Graphify recipes: onboard to a codebase, share a team graph, connect coding agents, load Neo4j and Postgres, and add graph impact to CI. +- [Graphify cookbooks: step-by-step recipes](https://docs.graphify.com/cookbooks/overview.md): Step-by-step Graphify recipes: onboard to a codebase, fix PR review findings, share a team graph, connect coding agents, and add graph impact to CI. - [How to understand a new codebase in 30 minutes](https://docs.graphify.com/cookbooks/onboard-to-a-codebase.md): Understand a new codebase in 30 minutes: build a knowledge graph, read its report and hub nodes, then ask questions that cite real files and lines. - [Plan a safe refactor with change impact analysis](https://docs.graphify.com/cookbooks/plan-a-refactor.md): Run change impact analysis before a refactor: list the callers of a function, trace paths between modules and find the tests to run, with Graphify. - [Find dead and duplicate code in your repos](https://docs.graphify.com/cookbooks/read-your-graph.md): Turn your code graph into a cleanup plan: read code areas, hubs and change impact, find dead and duplicate code, then hand it to your agent safely. diff --git a/platform/overview.mdx b/platform/overview.mdx index 0effc49..a2db724 100644 --- a/platform/overview.mdx +++ b/platform/overview.mdx @@ -8,7 +8,7 @@ hideFooterPagination: true icon: "house" --- -Graphify is the hosted product at app.graphify.com. We run it for you, so there's nothing to deploy. You use it through the console, the web UI at [app.graphify.com](https://app.graphify.com?utm_source=docs&utm_medium=link&utm_campaign=platform-overview). Your work lives in a workspace: a personal one for you, or a team workspace you share. Graphify brings repository graphs, wiki, pull request reviews, formal verification, Learning, and agent memory into that workspace for your team and coding assistants. +Graphify is shared codebase context for teams and coding agents, hosted at app.graphify.com. We run it for you, so there's nothing to deploy. You use it through the console, the web UI at [app.graphify.com](https://app.graphify.com?utm_source=docs&utm_medium=link&utm_campaign=platform-overview). Your work lives in a workspace: a personal one for you, or a team workspace you share. Graphify brings repository graphs, wiki, [cross-repo PR review](/platform/reviews), formal verification, Learning, and agent memory into that workspace for your team and coding assistants. New to Graphify? [Graphify in 5 minutes](/platform/quickstart) takes you from sign in to asking your coding agent about an indexed repository. @@ -29,6 +29,9 @@ To keep everything on your machine instead, use open-source Graphify, the `graph Use **Connect agent** in the console, or follow the [client-specific MCP instructions](https://graphify.com/mcp?utm_source=docs&utm_medium=link&utm_campaign=platform-overview). The hosted connection uses an OAuth sign-in so your assistant can access your Graphify workspace. + + **PR reviews** is on for a connected repository unless someone turns it off. Open a pull request, or mark a draft ready for review, and Graphify reviews it with the graph. On GitHub, the review shows up as a review comment and a **Graphify** check run. See [Review a pull request](/concepts/review-a-pull-request). + ## Start with your goal @@ -63,6 +66,7 @@ Structure and pattern discovery are rolling out separately and require a compati ## Common tasks - [Trace who calls this](/platform/trace): find the callers of a symbol and what a change can reach. +- [Review a pull request](/concepts/review-a-pull-request): get a review, run it again with `@graphify review`, and tell Graphify which findings were right. - [Read a review and a verdict](/platform/read-a-result): what a finding and a verification result each mean. - [When something fails](/platform/support/faqs): an empty graph, a failed build, or a plan limit. diff --git a/platform/quickstart.mdx b/platform/quickstart.mdx index 45cf9f3..4b2d6c4 100644 --- a/platform/quickstart.mdx +++ b/platform/quickstart.mdx @@ -116,7 +116,7 @@ To confirm your agent is connected and picks Graphify over grep, see [Check your Every page in app.graphify.com, explained. - Get graph-aware findings on your pull requests. + Get a cross-repo PR review on each new pull request, with findings that read the graph. Move from a local graph to Graphify. diff --git a/platform/reviews.mdx b/platform/reviews.mdx index 21fc961..b594743 100644 --- a/platform/reviews.mdx +++ b/platform/reviews.mdx @@ -17,6 +17,16 @@ Graphify reviews a pull request when it is opened, reopened, or marked ready for After new commits, a repository owner, member, or collaborator comments `@graphify review` on the pull request. Each review analyzes the full pull request, from where it branched to its latest commit. Reviews use the graph to understand how the changes affect the surrounding code. Reviews compare the pull request's changes against the indexed default branch. Nothing from the pull request's branch is saved as an index. [Repositories and builds](/platform/repositories) explains which branch the graph follows. +## Cross-repo impact + +When a repository is part of a cross-repository graph, its pull request reviews also load the latest graphs of the other repositories in that graph. Graphify links those repositories through the types they share and the calls between them. + +The review comment then includes a **Cross-repo impact** section. It says "This change impacts N symbol(s) across M peer repo(s)" and lists each affected repository with up to 10 of its symbols. It covers callers and dependents in the other repositories. If nothing is affected, the section says no cross-repo impact was detected among the indexed peer repositories. It also warns you when some repositories in the group aren't indexed yet. + +The section is advisory. It never changes whether the **Graphify** check passes or fails. + +There's nothing to turn on. To set it up, open **Graphs** > **Cross-Repository** in the app and select **Combine repositories**. A combined graph needs at least two repositories. See [Combine repositories](/platform/graphs#combine-repositories). + ## Know when a review is ready A review shows up in two places on the pull request: a review comment from the Graphify app, and a check run named **Graphify** on the reviewed commit. Formal verification posts its own check run, **Graphify Formal Verification**, and runs on every push. @@ -77,8 +87,8 @@ Expand the repository row to see **Unreferenced code (heuristic)**. It is not a When Graphify cannot start a review because of a plan limit, the skip is visible on the pull request's **Graphify** check (and as a notice on **Reviews** when the workspace is over its review allowance). Examples: -- Monthly or daily review limit reached — the check says reviews resume at the next UTC reset, and links to billing. -- Usage limit reached — same idea for the workspace usage cap. +- Monthly or daily review limit reached: the check says reviews resume at the next UTC reset, and links to billing. +- Usage limit reached: same idea for the workspace usage cap. When the workspace already has its maximum number of reviews running, a new review **waits in line** instead of being dropped. The check shows **Queued** with how many reviews are already running. If it does not start within the queue age limit, it is dropped; comment `@graphify review` to retry. Concurrent review limits by plan are on [Plans and pricing](/platform/support/pricing#build-concurrency). diff --git a/scripts/build-llms-txt.py b/scripts/build-llms-txt.py index ab63eb0..315599f 100644 --- a/scripts/build-llms-txt.py +++ b/scripts/build-llms-txt.py @@ -20,6 +20,7 @@ - Setting Graphify up for a user? Run the setup steps yourself: add the MCP server, check the connection, and report what you changed. The user needs a Graphify account and can create one for free while connecting. Ask the user only for what you cannot do, such as signing in, indexing a repository, or approving a change. Follow https://docs.graphify.com/for-agents.md step by step. - Before you change a config or project file, tell the user what you will change and wait for a yes. - Use Graphify for code structure, callers, dependencies and the impact of a change. Use grep for exact strings, such as error messages or config keys. +- Graphify reviews a pull request when it is opened, reopened or marked ready for review, and `@graphify review` runs it again after new commits. Read results with the review_status tool, which does not start a review. See https://docs.graphify.com/platform/reviews.md. - Use only commands, flags and endpoints shown on these pages. Do not invent one. - Add .md to any page URL to read it as Markdown. Prefer this index and single .md pages. https://docs.graphify.com/llms-full.txt has every page in one file, but it is about 850 KB, so most agents only see the start of it. - Search these docs over MCP, with no sign-in: https://docs.graphify.com/mcp. It is a different server from the product server at https://api.graphify.com/mcp. diff --git a/snippets/try-cloud-card.mdx b/snippets/try-cloud-card.mdx index 6d817c3..4f5e7af 100644 --- a/snippets/try-cloud-card.mdx +++ b/snippets/try-cloud-card.mdx @@ -1,3 +1,3 @@ - We build and host the graph for you. The Free plan (up to 3 repositories) needs no card. Pro and Teams start with a 14-day trial, card required. + We build and host the graph for you, and review your pull requests with it. The Free plan (up to 3 repositories) needs no card. Pro and Teams start with a 14-day trial, card required.