Unabyss

By OneType P.S.A.

Get New App Alerts

Open in ChatGPT

Open in Claude

Overview

Unabyss is a user-owned context layer for AI tools. It connects to the sources where your context already lives — Gmail, Google Drive, Calendar, Notion, Slack, GitHub, LinkedIn, X, Obsidian — and extracts a structured vault covering who you are, what you're working on, and how you communicate. With the connector enabled, Claude can pull that context at the start of any conversation instead of you re-explaining your role, company, stack, and preferences every time. What you can do: - Look up specific facts about yourself, your company, or past decisions, and ask synthesis questions across your entire context - Save durable facts, decisions, and preferences back to your vault as they come up in conversation - Load skills — ready-made playbooks for recurring work like daily briefs, investor updates, pre-call research, and content drafting - Generate topic-scoped Markdown exports to reuse in tools that don't support MCP - Connect new sources directly from chat, without leaving the conversation Your context stays yours. Permissions are granular per tool and per layer (Identity / Profile / Mind), connected sources keep context self-updating, and access can be revoked at any time. Unabyss is MCP-native and cross-platform, so the same vault works in Claude, Claude Code, Cursor, and any other MCP client.

Tools

agentic_query

ChatGPT

Deep reasoning over the user's full context -- synthesis, comparison, or reflective analysis. If processing exceeds 30s, returns a query_id for polling via agentic_query_read.

agentic_query_read

ChatGPT

Poll for a pending agentic_query result by query_id. Returns the answer if ready, or pending status.

export_create

ChatGPT

Create a new export on a topic. Multi-minute background generation -- only on explicit user request. Accepts either a free-text topic or a preset_slug (not both).

export_create_from_text

ChatGPT

Save text as an export without generation. Instant and free. Use when the user asks to save a response or summary.

export_list

ChatGPT

List the user's exports (topic-scoped markdown dossiers). Returns id, title, status, and a staleness flag.

export_read

ChatGPT

Read the full markdown content of a specific export by ID. The export must be in 'ready' status.

export_refresh

ChatGPT

Regenerate an existing export against the latest knowledge base. Multi-minute -- only on explicit user request.

forget

ChatGPT

Permanently forget specific memories. Requires explicit user confirmation before calling. Get memory_ids from the search tool (not query). Envelope IDs come from store's changed sources / search source_id. Pass either memory_ids (point IDs) or envelope_ids (source IDs). Exactly one must be provided. Caps: 50 memory_ids / 20 envelope_ids. Forgotten facts cannot resurface even after re-import. Fact-level forget hides the fact from search and query, but archive text can still appear in verbatim hits and exports until the envelope is forgotten. Does not purge the identity summary (whoami).

get_skill

ChatGPT

Load a skill by slug (as returned by list_skills). With type 'markdown' (default) returns the playbook body plus any supporting-file paths; with type 'skill-file' returns a link to download the packaged .skill bundle (for clients that install skills by uploading the file).

list_integrations

ChatGPT

List the user's connected integrations and catalog apps they can still connect. Returns connected rows (one per live connection with sync metadata) and available rows (catalog apps not yet connected). Use when the user asks about integrations, connections, or data availability.

list_skills

ChatGPT

List available skills -- step-by-step playbooks for common tasks. Returns each skill's slug, name, and description. Consult this when a user task may match a skill, then load it with get_skill.

propose_connection

ChatGPT

Generate a one-click connect URL for a specific app. Use when the user wants to connect a new data source.

query

ChatGPT

Look up personal context. Use when the user references themselves, their work, preferences, or prior decisions. Returns a compiled text answer. For deep synthesis or reflective analysis, use agentic_query instead.

read_skill_file

ChatGPT

Read a skill's supporting file by slug and relative path (as listed in get_skill). Returns the file's text content.

read_source

ChatGPT

Read the full text of a live source. Pass provider plus the opaque source_id minted by search_sources. Do not pass a raw Google file, thread, event, or tl;dv meeting id, a Calendar day-agenda key, or a raw GitHub owner/repo/path or issue number. Providers: google_drive, gmail, google_calendar, tldv, github. Use after search_sources when the user needs the body or full transcript rather than a search hit.

search

ChatGPT

Semantic search over the user's knowledge base. Returns a JSON list of hits with memory_id, source_id/source_ids, score, fingerprint, and truncated chunk_text. Permission-filtered tokens cannot call search; use query (sanitized) instead. Use to locate memory_ids before forget (not query). Default limit 8.

search_sources

ChatGPT

Search live connected apps by keywords. Providers: google_drive, gmail, google_calendar, tldv, github. Drive uses keyword search (fullText contains), not Vectorize. Gmail uses Gmail search q (not the ingest promo/noreply filter). Calendar requires ISO time_min and time_max (max 31 days) and uses Google events.list q; empty Calendar query lists that window. tldv is a capped in-process title and date filter (not a search API): case-insensitive substring on meeting name, optional time_min/time_max on happenedAt, empty query lists the cap. github is the GitHub Search API over selected repos (code, issues, and pull requests); a non-empty query is required. Use after a memory miss, a named file/thread/event/meeting/issue/PR, recency, or full transcript/body. Optional account_id and limit (default 8, max 20). Returns hits with provider, kind, source_id, title, occurred_at, url, in_memory, plus truncated. No bodies. Then call read_source with source_id. If the app is not connected, call propose_connection. Do not scrape Drive, Gmail, Calendar, tl;dv, or GitHub URLs.

store

ChatGPT

Persist a durable fact about the user -- preference, decision, project, role, or tool choice. Stored for retrieval across AI tools.

unabyss_setup

ChatGPT

Optional, step-based setup for this connection. Call with no argument (or step 'step1') to offer the user setup, then advance in order: 'step2' to connect apps (uses list_integrations / propose_connection); 'step3' to ask what to save back to Unabyss; 'step4' with context_preference for the system prompt; then 'first-prompt' with instructions_installed to finish - it records setup as complete, unblocks every tool, and offers the user a first thing to run. Use 'completed' or 'skipped' only when the user stops early, and only after they explicitly choose - never decide that yourself (either way, future setup prompts stop on this connection). Skills are not part of the flow: only if the user asks for skills, 'step5' browses them and 'all-skills', 'essential-skills', or '<category>-skills' (e.g. 'sales-skills') list them with install instructions.