Imports coding conversations from GitHub Copilot, distills them into structured notes with semantic search and knowledge graph, and provides MCP tools for agents to reuse insights.
Local-first AI PKM for coding conversations
Website · Download Desktop · npm · Docs · 简体中文
ChatCrystal is a local-first AI PKM app for developers who solve real problems with Claude Code, Cursor, Codex CLI, Trae, and GitHub Copilot.
It turns scattered AI coding conversations into structured notes, semantic search, a tag knowledge graph, Markdown exports, and MCP memory your agents can reuse. If this fits your workflow, a star helps more builders find a private, local-first way to keep their AI work memory.
Download the latest Windows installer from GitHub Releases. After installing, launch ChatCrystal, configure your LLM and embedding providers in Settings, then click Import.
npm install -g chatcrystal
crystal serve -d
crystal importThen open http://localhost:3721 in your browser.
Prefer self-hosting ChatCrystal for multiple devices? See Docker Cloud Deployment after the product overview.
Related MCP server: decisionnode/DecisionNode
Imports AI coding conversations from local tool data directories.
Distills conversations into structured notes with titles, summaries, conclusions, snippets, and tags.
Searches knowledge semantically with embeddings and relation-aware result expansion.
Builds a tag knowledge graph where knowledge points are tags and edges show normalized co-occurrence.
Exposes CLI and MCP tools so agents can recall and write back reusable experience.
Runs locally with configurable LLM and embedding providers.
crystal status # Server status and DB stats
crystal import [--source claude-code] # Scan and import conversations
crystal search "query" [--limit 10] # Semantic search
crystal notes list [--tag X] # Browse notes
crystal notes get <id> # View note detail
crystal summarize --all # Batch summarize
crystal config get # View config
crystal serve -d # Start server in background
crystal serve stop # Stop background server
crystal mcp # Start MCP stdio serverTopic | English | 简体中文 |
User guide | ||
Development | ||
MCP and agents | ||
Experience quality gate | ||
Agent skills |
Node.js >= 20
An LLM provider for summarization
An embedding provider for semantic search
LLM and embedding providers are configured separately. Large language models such as Claude, GPT, and Qwen are not embedding models. See the user guide for provider examples.
git clone https://github.com/ZengLiangYi/ChatCrystal.git
cd ChatCrystal
npm install
npm run devDevelopment server ports:
API/server: http://localhost:3721
Vite client: http://localhost:13721
See docs/DEVELOPMENT.md for architecture, testing, build, and release details.
The default Compose deployment runs only the ChatCrystal service. It stores data in the chatcrystal-data volume mounted at /data inside the container.
git clone https://github.com/ZengLiangYi/ChatCrystal.git
cd ChatCrystal
docker compose up -dThe default docker-compose.yml pulls ghcr.io/zengliangyi/chatcrystal:latest from GitHub Container Registry. Set CHATCRYSTAL_IMAGE_TAG to pin a published version. To build from source instead, run docker compose -f docker-compose.yml -f docker-compose.build.yml up -d --build.
To update an existing Docker deployment, run docker compose pull && docker compose up -d. Maintainers only: after the first GHCR publish, make the ghcr.io/zengliangyi/chatcrystal package public in GitHub Packages; the release workflow verifies anonymous pull access before passing.
Compose binds ChatCrystal to 0.0.0.0:3721 by default so other devices can reach the cloud core through the host IP. Set CHATCRYSTAL_HOST_PORT to change the host port, or set BIND_ADDRESS=127.0.0.1 when a local-only reverse proxy fronts it. For public cloud access, an HTTPS reverse proxy is recommended for safer token transport.
On Windows Docker Desktop, a published port may still need extra host networking configuration before it is reachable through the host LAN IP. For cloud-mode testing, verify http://<host-ip>:<host-port>/api/health from the client device first; if it cannot connect, configure Windows port forwarding/firewall rules or deploy the cloud core on a real remote host.
On first start without CHATCRYSTAL_API_TOKEN, open the Web UI and enter the setup code printed in container logs or stored at /data/setup-code, then choose one shared API token for your devices.
To rotate or reset the Docker cloud token:
# If you still know the current token, rotate it online.
crystal --base-url https://chatcrystal.example.com token rotate "new-long-token-at-least-16-chars" --current "old-token"
crystal connect https://chatcrystal.example.com --token "new-long-token-at-least-16-chars"
# If you forgot the token and did not set CHATCRYSTAL_API_TOKEN, reset stored auth in the container.
docker compose exec chatcrystal crystal token reset --yes
docker compose logs chatcrystal --tail=80
docker compose exec chatcrystal cat /data/setup-codeIf your deployment sets CHATCRYSTAL_API_TOKEN, that environment variable is the active token source. Change it in your .env or Compose environment and recreate the container with docker compose up -d --force-recreate.
To use an existing Ollama or external API, configure provider URLs in the Web UI or environment. In Docker, localhost means inside the container; use CHATCRYSTAL_DOCKER_LLM_BASE_URL and CHATCRYSTAL_DOCKER_EMBEDDING_BASE_URL for Compose-time provider URL overrides. Docker Desktop can reach host Ollama at http://host.docker.internal:11434, or you can use a remote HTTPS/OpenAI-compatible API.
Install or run the CLI on the device that has Claude Code, Cursor, Codex CLI, Trae, or GitHub Copilot history:
crystal connect https://chatcrystal.example.com --token "your-long-token"
crystal import --yesThe CLI scans local histories, parses them locally, and uploads normalized conversations to the cloud. The cloud never scans your local filesystem. Imported conversations are not summarized automatically; use the Web UI or crystal summarize --all when you are ready. HTTPS is recommended for cloud access, but HTTP works when that is the deployment you choose.
Get the full content of a note including title, summary, key conclusions, code snippets, and tags.
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | Note ID |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description indicates a read operation and lists included fields but lacks details on return format, permissions, or constraints. With no annotations provided, the description carries the burden for transparency, and while adequate for a simple read, it could be more explicit.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single sentence with no redundancy, but it could be slightly more structured (e.g., bullet points for output fields) without increasing verbosity. Still, it is concise and clear.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's simplicity (one parameter, no output schema, no annotations), the description adequately covers purpose and output. However, it does not mention return format or prerequisites, which would be beneficial for complete context.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100% (the only parameter 'id' has a description 'Note ID'). The tool description does not add additional meaning beyond the schema; therefore, baseline score of 3 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action ('Get') and the resource ('full content of a note'). It lists specific included elements (title, summary, key conclusions, code snippets, tags), making the tool's function distinct from siblings like list_notes or get_relations.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is provided on when to use this tool versus alternatives (e.g., list_notes for metadata or search_knowledge for searching). The agent must infer usage solely from the description and sibling names.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get related notes for a given note, including relationship type and confidence score.
| Name | Required | Description | Default |
|---|---|---|---|
| noteId | Yes | Note ID to find relations for |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must carry behavioral disclosure. It only states what is returned but does not mention safety (e.g., read-only), authentication needs, or rate limits.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, clear sentence with no redundant words. It is appropriately sized for the task.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With one parameter and no output schema, the description is minimal. It states the output includes relationship type and confidence score but does not specify if results are paginated, ordered, or empty-possible. Adequate for a simple tool but lacks completeness.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% for the single parameter noteId. The schema already describes it as 'Note ID to find relations for'. The description adds no additional meaning beyond that.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'Get', the resource 'related notes', and specifies the output includes 'relationship type and confidence score'. It distinguishes from sibling tools like get_note and list_notes by focusing on relationships.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage when needing related notes for a given note, but does not provide explicit when-not or alternative tools. No guidance on filtering or prerequisites.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List note summaries for browsing and narrowing the ChatCrystal knowledge base. Use this when you need paginated notes filtered by tag or title/summary keyword. Use search_knowledge instead for semantic relevance ranking, and get_note when you already have an id and need the full note body. Returns note metadata and summaries, not full note content.
| Name | Required | Description | Default |
|---|---|---|---|
| tag | No | Exact tag name to filter notes by, for example "mcp" or "cursor". | |
| page | No | 1-based page number for paginated note summaries. Each page returns up to 20 notes. | |
| search | No | Literal keyword filter applied to note title and summary; not semantic search. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It discloses output is metadata/summaries not full content, and implies read-only nature. Missing info on auth or rate limits but adequate for this simple tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences, front-loaded with purpose, no wasted words. Efficient and structured.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given simple tool with 3 params, no output schema, description fully covers what the tool does, when to use, parameter roles, and output limitations.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. Description adds context about pagination and search being literal, but much is already in schema descriptions. No significant extra meaning beyond schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Description clearly states it lists note summaries for browsing, and explicitly distinguishes from sibling tools search_knowledge and get_note. Specific verb-resource pair.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides explicit when to use (browsing, filtering by tag or keyword) and when not (semantic search or full note retrieval), naming alternatives.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Retrieve reusable task memories before starting substantive coding work. Use this at the beginning of implementation, debugging, migration, configuration, investigation, refactor, or optimization tasks to load project-scoped memories first and optional global lessons second. Use mode="debug" when the user reports an error, failing command, regression, or incident; include error_signatures and related_files when available. Use search_knowledge instead for ad hoc semantic note search that is not tied to the current task. This tool is read-only and returns ranked memories plus optional related-note context without writing anything.
| Name | Required | Description | Default |
|---|---|---|---|
| mode | No | Use task for normal work and debug when the task starts from an error, failing test, or incident. | task |
| task | Yes | Current task context used to retrieve relevant project and global memories. | |
| options | No | Optional limits and relation expansion controls for recall results. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It states the tool is read-only, returns ranked memories plus optional related-note context, and writes nothing. This sufficiently discloses behavioral traits without contradictions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Single paragraph that is front-loaded with purpose and well-structured. Every sentence adds value; no wasted words. Efficiently covers usage, mode, and exclusions.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given complexity (3 parameters, nested object, no output schema), description explains return values (ranked memories plus optional context) and main use cases. Lacks details on exact output format but sufficient for agent understanding.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so baseline is 3. The description adds some context (e.g., use debug mode for errors) but mainly restates schema information. No significant additional semantics beyond schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool retrieves reusable task memories before starting substantive coding work, with specific verb ('Retrieve') and resource ('reusable task memories'). It distinguishes from sibling tool search_knowledge by specifying when to use that alternative.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides explicit guidance on when to use (at beginning of implementation, debugging, etc.) and when not to (ad hoc note search via search_knowledge). Also gives mode-specific instructions (debug mode for errors) and lists optional parameters for context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Semantic search across your AI conversation knowledge base. Returns matching notes ranked by relevance.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | Maximum number of results | |
| query | Yes | Search query text |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description mentions it returns notes ranked by relevance, indicating a read and search behavior, but lacks details on pagination, result format, no-result behavior, or any side effects. With no annotations, more disclosure would be beneficial.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise with two sentences, front-loading the core purpose. There is no unnecessary text, making it efficient for an AI agent to parse.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple tool with two parameters and no output schema, the description covers the main functionality but omits details on the return format of notes and edge cases. It is adequate but not fully comprehensive.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema already provides descriptions for both parameters (query and limit). The tool description adds no additional meaning beyond what the schema states, thus baseline score of 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool performs semantic search across a knowledge base and returns ranked notes. It distinguishes itself from siblings like list_notes (general listing) and get_note (retrieving a specific note) by specifying the search nature and ranking.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is provided on when to use this tool versus alternatives such as list_notes or recall_for_task. The description only states what it does without contextual recommendations.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Dry-run validation for a candidate task memory before calling write_task_memory. Use this after meaningful work and before persisting a lesson to check whether the candidate is durable, specific, reusable, and shaped like a high-quality ChatCrystal note. It has no side effects and never writes to the knowledge base. Returns acceptance, rejection reason, warnings, and materialized note fields so agents can revise the candidate or skip weak work logs.
| Name | Required | Description | Default |
|---|---|---|---|
| mode | Yes | Use auto for agent-generated writebacks and manual for explicit user-curated memories. | |
| task | Yes | Current task context used to scope, rank, and store memories. | |
| scope | No | Store as project memory by default; global is reserved for broadly reusable manual lessons. | |
| memory | Yes | Candidate ChatCrystal note content to validate or persist as reusable task memory. | |
| source_run_key | No | Idempotency key for auto writebacks; required in auto mode to avoid duplicate memory receipts. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description fully discloses behavior: 'Dry-run validation', 'has no side effects', 'never writes to the knowledge base', and describes return values (acceptance, rejection reason, warnings, materialized note fields). No contradictions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences efficiently cover purpose, usage, and behavior without fluff. Front-loaded with primary action, then usage guidance, then behavior and returns. Every sentence adds value.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given moderate complexity (5 params, nested objects) and no output schema, the description covers return fields (acceptance, rejection reason, warnings, materialized note fields) and states the tool is for validation. Could elaborate more on what constitutes 'high-quality' criteria but is sufficient for an agent to use correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. The description adds context about what the tool does overall but does not discuss individual parameters beyond what the schema already provides. It does not compensate for any missing schema details.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool validates a candidate task memory before persisting, using specific verbs ('validate') and resources ('task memory'). It distinguishes itself from sibling write_task_memory with the 'Dry-run' qualifier and explicitly names it as the counterpart.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides clear context: 'Use this after meaningful work and before persisting a lesson'. States it has no side effects, indicating safe dry-run usage. However, does not explicitly state when not to use or list alternatives beyond implying write_task_memory is the next step.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Persist a task memory only when it can become a high-quality ChatCrystal note: specific title, concrete summary, meaningful key conclusions, and a durable reusable lesson such as a pitfall, fix, decision, pattern, or symptom-to-resolution mapping. Do not write one-time environment checks, version/status reports, ordinary progress logs, or vague robustness claims. Weak auto writebacks are skipped by core validation and recorded only as receipts.
| Name | Required | Description | Default |
|---|---|---|---|
| mode | Yes | Use auto for agent-generated writebacks and manual for explicit user-curated memories. | |
| task | Yes | Current task context used to scope, rank, and store memories. | |
| scope | No | Store as project memory by default; global is reserved for broadly reusable manual lessons. | |
| memory | Yes | Candidate ChatCrystal note content to validate or persist as reusable task memory. | |
| source_run_key | No | Idempotency key for auto writebacks; required in auto mode to avoid duplicate memory receipts. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Discloses quality validation and skip behavior for weak entries, but with no annotations, it misses details like side effects, auth requirements, rate limits, or error handling. The description carries the full burden due to missing annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Description is a single paragraph that is front-loaded with purpose and criteria. It includes examples and exclusions without excessive verbosity. Could be slightly more concise but effectively conveys key information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the complex nested parameters and no output schema, the description covers purpose, usage conditions, and quality expectations. It lacks explanation of return values (receipts) but otherwise is fairly complete.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so parameters are well-documented in the schema. The description adds context on quality criteria but does not enhance parameter meaning beyond what the schema provides. Baseline 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool persists high-quality task memories as ChatCrystal notes with specific criteria (title, summary, conclusions, reusable lesson). It distinguishes from weak auto writebacks but does not explicitly differentiate from sibling tools like validate_task_memory or recall_for_task.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides explicit when-to-use (only for high-quality, durable lessons) and when-not-to-use (one-time checks, status reports, vague claims). Mentions that weak writebacks are skipped. Lacks explicit alternatives or comparison to sibling tools.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.5.6list_notes3 fields changedInput schema / properties / page / descriptionPrevious value: -"Page number"New value: +"1-based page number for paginated note summaries. Each page returns up to 20 notes."Input schema / properties / search / descriptionPrevious value: -"Filter by keyword in title/summary"New value: +"Literal keyword filter applied to note title and summary; not semantic search."Input schema / properties / tag / descriptionPrevious value: -"Filter by tag name"New value: +"Exact tag name to filter notes by, for example \"mcp\" or \"cursor\"."recall_for_task16 fields changedInput schema / properties / mode / descriptionAdded value: +"Use task for normal work and debug when the task starts from an error, failing test, or incident."Input schema / properties / options / descriptionAdded value: +"Optional limits and relation expansion controls for recall results."Input schema / properties / options / properties / global_limit / descriptionAdded value: +"Maximum number of cross-project/global lessons to append after project memories."Input schema / properties / options / properties / include_relations / descriptionAdded value: +"Whether to include related-note context for returned memories."Input schema / properties / options / properties / project_limit / descriptionAdded value: +"Maximum number of project-scoped memories to return first."Input schema / properties / task / descriptionAdded value: +"Current task context used to retrieve relevant project and global memories."Input schema / properties / task / properties / branch / descriptionAdded value: +"Current VCS branch when relevant to the task."Input schema / properties / task / properties / cwd / descriptionAdded value: +"Current working directory of the agent session."Input schema / properties / task / properties / error_signatures / descriptionAdded value: +"Concrete errors, stack traces, failing commands, or symptoms. Most useful with debug tasks."Input schema / properties / task / properties / files_touched / descriptionAdded value: +"Files already touched or expected to be touched; improves project memory matching."Input schema / properties / task / properties / goal / descriptionAdded value: +"Plain-language task goal or user request. Include enough context to retrieve relevant memories."Input schema / properties / task / properties / project_dir / descriptionAdded value: +"Absolute project directory when known; helps ChatCrystal match memories to the right local workspace."Input schema / properties / task / properties / project_key / descriptionAdded value: +"Stable project identifier used to prioritize project-scoped memories, such as a repository or workspace key."Input schema / properties / task / properties / related_files / descriptionAdded value: +"Additional files related to the task but not necessarily modified."Input schema / properties / task / properties / source_agent / descriptionAdded value: +"AI coding tool or agent that is calling ChatCrystal; use unknown when unsure."Input schema / properties / task / properties / task_kind / descriptionAdded value: +"Kind of work being performed. Use debug for failures; choose the closest non-debug category for planned work."validate_task_memory30 fields changedInput schema / properties / memory / descriptionAdded value: +"Candidate ChatCrystal note content to validate or persist as reusable task memory."Input schema / properties / memory / properties / code_snippets / descriptionAdded value: +"Small snippets that make the memory actionable without copying large files."Input schema / properties / memory / properties / code_snippets / items / properties / code / descriptionAdded value: +"Minimal code, command, config, or query that illustrates the reusable lesson."Input schema / properties / memory / properties / code_snippets / items / properties / description / descriptionAdded value: +"Why this snippet matters and when to reuse it."Input schema / properties / memory / properties / code_snippets / items / properties / language / descriptionAdded value: +"Programming or markup language for the snippet."Input schema / properties / memory / properties / decisions / descriptionAdded value: +"Durable design, product, architecture, or process decisions made during the task."Input schema / properties / memory / properties / error_signatures / descriptionAdded value: +"Exact errors or symptoms that should trigger this memory in future debug recall."Input schema / properties / memory / properties / files_touched / descriptionAdded value: +"Files that provide useful provenance for the memory."Input schema / properties / memory / properties / key_conclusions / descriptionAdded value: +"Important takeaways that should be recalled before similar future work."Input schema / properties / memory / properties / outcome_type / descriptionAdded value: +"Primary kind of reusable memory being saved."Input schema / properties / memory / properties / pitfalls / descriptionAdded value: +"Mistakes, traps, or failure modes future agents should avoid."Input schema / properties / memory / properties / resolution / descriptionAdded value: +"Specific fix or action that resolved the issue."Input schema / properties / memory / properties / reusable_patterns / descriptionAdded value: +"Generalizable implementation, debugging, migration, or configuration patterns."Input schema / properties / memory / properties / root_cause / descriptionAdded value: +"Underlying cause of the problem when the memory is about a fix or pitfall."Input schema / properties / memory / properties / summary / descriptionAdded value: +"Concrete summary of what was learned or decided, written so it remains useful in a later session."Input schema / properties / memory / properties / tags / descriptionAdded value: +"Short tags for retrieval, such as framework, subsystem, source tool, or failure type."Input schema / properties / memory / properties / title / descriptionAdded value: +"Specific note title. Prefer the durable lesson over a generic task name."Input schema / properties / mode / descriptionAdded value: +"Use auto for agent-generated writebacks and manual for explicit user-curated memories."Input schema / properties / scope / descriptionAdded value: +"Store as project memory by default; global is reserved for broadly reusable manual lessons."Input schema / properties / source_run_key / descriptionAdded value: +"Idempotency key for auto writebacks; required in auto mode to avoid duplicate memory receipts."Input schema / properties / task / descriptionAdded value: +"Current task context used to scope, rank, and store memories."Input schema / properties / task / properties / branch / descriptionAdded value: +"Current VCS branch when relevant to the task."Input schema / properties / task / properties / cwd / descriptionAdded value: +"Current working directory of the agent session."Input schema / properties / task / properties / error_signatures / descriptionAdded value: +"Concrete errors, stack traces, failing commands, or symptoms. Most useful with debug tasks."Input schema / properties / task / properties / files_touched / descriptionAdded value: +"Files already touched or expected to be touched; improves project memory matching."Input schema / properties / task / properties / goal / descriptionAdded value: +"Plain-language task goal or user request. Include enough context to retrieve relevant memories."Input schema / properties / task / properties / project_dir / descriptionAdded value: +"Absolute project directory when known; helps ChatCrystal match memories to the right local workspace."Input schema / properties / task / properties / project_key / descriptionAdded value: +"Stable project identifier used to prioritize project-scoped memories, such as a repository or workspace key."Input schema / properties / task / properties / source_agent / descriptionAdded value: +"AI coding tool or agent that is calling ChatCrystal; use unknown when unsure."Input schema / properties / task / properties / task_kind / descriptionAdded value: +"Kind of work being performed. Use debug for failures; choose the closest non-debug category for planned work."write_task_memory30 fields changedInput schema / properties / memory / descriptionAdded value: +"Candidate ChatCrystal note content to validate or persist as reusable task memory."Input schema / properties / memory / properties / code_snippets / descriptionAdded value: +"Small snippets that make the memory actionable without copying large files."Input schema / properties / memory / properties / code_snippets / items / properties / code / descriptionAdded value: +"Minimal code, command, config, or query that illustrates the reusable lesson."Input schema / properties / memory / properties / code_snippets / items / properties / description / descriptionAdded value: +"Why this snippet matters and when to reuse it."Input schema / properties / memory / properties / code_snippets / items / properties / language / descriptionAdded value: +"Programming or markup language for the snippet."Input schema / properties / memory / properties / decisions / descriptionAdded value: +"Durable design, product, architecture, or process decisions made during the task."Input schema / properties / memory / properties / error_signatures / descriptionAdded value: +"Exact errors or symptoms that should trigger this memory in future debug recall."Input schema / properties / memory / properties / files_touched / descriptionAdded value: +"Files that provide useful provenance for the memory."Input schema / properties / memory / properties / key_conclusions / descriptionAdded value: +"Important takeaways that should be recalled before similar future work."Input schema / properties / memory / properties / outcome_type / descriptionAdded value: +"Primary kind of reusable memory being saved."Input schema / properties / memory / properties / pitfalls / descriptionAdded value: +"Mistakes, traps, or failure modes future agents should avoid."Input schema / properties / memory / properties / resolution / descriptionAdded value: +"Specific fix or action that resolved the issue."Input schema / properties / memory / properties / reusable_patterns / descriptionAdded value: +"Generalizable implementation, debugging, migration, or configuration patterns."Input schema / properties / memory / properties / root_cause / descriptionAdded value: +"Underlying cause of the problem when the memory is about a fix or pitfall."Input schema / properties / memory / properties / summary / descriptionAdded value: +"Concrete summary of what was learned or decided, written so it remains useful in a later session."Input schema / properties / memory / properties / tags / descriptionAdded value: +"Short tags for retrieval, such as framework, subsystem, source tool, or failure type."Input schema / properties / memory / properties / title / descriptionAdded value: +"Specific note title. Prefer the durable lesson over a generic task name."Input schema / properties / mode / descriptionAdded value: +"Use auto for agent-generated writebacks and manual for explicit user-curated memories."Input schema / properties / scope / descriptionAdded value: +"Store as project memory by default; global is reserved for broadly reusable manual lessons."Input schema / properties / source_run_key / descriptionAdded value: +"Idempotency key for auto writebacks; required in auto mode to avoid duplicate memory receipts."Input schema / properties / task / descriptionAdded value: +"Current task context used to scope, rank, and store memories."Input schema / properties / task / properties / branch / descriptionAdded value: +"Current VCS branch when relevant to the task."Input schema / properties / task / properties / cwd / descriptionAdded value: +"Current working directory of the agent session."Input schema / properties / task / properties / error_signatures / descriptionAdded value: +"Concrete errors, stack traces, failing commands, or symptoms. Most useful with debug tasks."Input schema / properties / task / properties / files_touched / descriptionAdded value: +"Files already touched or expected to be touched; improves project memory matching."Input schema / properties / task / properties / goal / descriptionAdded value: +"Plain-language task goal or user request. Include enough context to retrieve relevant memories."Input schema / properties / task / properties / project_dir / descriptionAdded value: +"Absolute project directory when known; helps ChatCrystal match memories to the right local workspace."Input schema / properties / task / properties / project_key / descriptionAdded value: +"Stable project identifier used to prioritize project-scoped memories, such as a repository or workspace key."Input schema / properties / task / properties / source_agent / descriptionAdded value: +"AI coding tool or agent that is calling ChatCrystal; use unknown when unsure."Input schema / properties / task / properties / task_kind / descriptionAdded value: +"Kind of work being performed. Use debug for failures; choose the closest non-debug category for planned work."v0.5.5get_noteget_relationslist_notesrecall_for_tasksearch_knowledgevalidate_task_memorywrite_task_memoryEach tool serves a clearly distinct purpose: getting full notes, listing summaries, semantic search, task-specific recall, relation retrieval, validation, and writing. No overlapping functionality.
All tool names follow a consistent verb_noun pattern using lowercase and underscores, e.g., get_note, list_notes, write_task_memory, ensuring predictability.
With 7 tools, the set is well-scoped for a knowledge base server, covering creation, retrieval, search, and validation without being overwhelming or sparse.
The tool surface covers core operations but lacks update and delete capabilities for notes, which may create dead ends if an agent needs to modify or remove a memory.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Private persistent memory for Claude, ChatGPT & Gemini via MCP - semantic search, zero-code setup.
Portable memory for AI agents: capture once, recall across Claude, Cursor, and any MCP client.
Persistent AI memory shared across Claude, ChatGPT, coding agents, and compatible MCP clients.
Persistent memory for AI assistants. Save once; recall from Claude, ChatGPT, or any MCP client.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/ZengLiangYi/ChatCrystal'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
