PowerContext v0.0.1

PowerContext v0.0.1

PowerContext is PowerMem 2.0, the upgraded version of PowerMem. It turns human-agent work into durable,
project-scoped context so that a new agent session can recover decisions, outcomes, current state, and next steps
without relying on chat history alone.

This initial release establishes the first end-to-end PowerContext workflow: capture useful work context, store it
with provenance, recall it within a bounded project scope, and hand work over to another session or agent.

Highlights

• Durable project context: remember, search, revise, and retire project-scoped Memory entries with stable
  citations and revision history.
• Agent handoffs: preserve outcomes, current state, disposition, evidence, and next actions so work can continue
  across sessions.
• Codex integration: install the PowerContext Codex plugin to recall relevant context and perform explicit Memory
  operations while coding. Server failures fail open and do not block normal Codex work.
• Local-first setup: start with a persistent SQLite database and a loopback Server without configuring an
  inference provider. Optional inference enables extraction and generation workflows.
• Multiple integration surfaces: use the Core SDK, asynchronous Python client, CLI, HTTP API, or MCP tools.
• Governed context lifecycle: review Experience and Skill candidates before they become available for recall or
  export.
• Operations and diagnostics: health and readiness checks, a web dashboard, Prometheus-compatible metrics, and
  OpenTelemetry tracing support.

Install

PowerContext requires Python 3.11 or later and supports Python versions below 4.0. Install the CLI and local Server
with uv:

uv tool install "powercontext[cli,server]==0.0.1"

To connect PowerContext to Codex, install the plugin from the matching release tag:

powercontext setup codex --source oceanbase/powercontext --ref v0.0.1
powercontext server run

In another terminal, verify the installation:

powercontext doctor
powercontext doctor codex

For a Python application that only needs the asynchronous client:

uv add "powercontext[client]==0.0.1"

Important Notes

• v0.0.1 is the first PowerContext package release. Public APIs, configuration, and persisted schemas may evolve
  as the project moves beyond its initial release.
• The default Server binds to 127.0.0.1:8000 and stores data in the operating system's user data directory.
• Explicit Memory operations work without an inference provider. Model-backed Memory extraction, reranking,
  Experience or Skill generation, and generated Handoff workflows require the corresponding inference configuration.
• Treat recalled context as untrusted evidence, and do not store credentials or other secrets in Memory.

Learn More

• PowerContext repository
• Codex quickstart
• Configuration reference
• Report an issue

Thank you for trying PowerContext. We welcome feedback on installation, agent handoffs, context quality, and the
developer experience.

v1.1.7

What's Changed

• fix(ci): dispatch publish workflow without git checkout by @wayyoungboy in #1116
• feat(plugin): enable transcript inference by default; remove recall/remember skills by @lightzt99 in #1126
• Fix powermem-mcp help to exit without starting the server by @wayyoungboy in #1124
• fix(ci): avoid duplicate generated release notes by @wayyoungboy in #1123
• chore: bump version to 1.1.7 by @wayyoungboy in #1127

Full Changelog: v1.1.6...v1.1.7

PowerMem 1.1.6

What's Changed

• feat: sqlite default to coding agent by @lightzt99 in #1031
• docs: update LOCOMO benchmark results by @wayyoungboy in #1089
• Add Chinese docs and website localization by @wayyoungboy in #1090
• fix(mcp): preserve Snowflake ID precision across MCP tool boundary by @lightzt99 in #1100
• feat: use uv instead of bare python command by @yihong0618 in #1106
• feat(init): AskUserQuestion interactive flow + remote-init mode + dashboard 0.0.0.0 by @lightzt99 in #1101
• fix(scripts): preserve doc dirs in sync-website-docs (Fixes #1096) by @kic635 in #1097
• fix(server): push run_id filter to storage for session timeline by @kic635 in #1099
• Publish individual release binary assets by @wayyoungboy in #1108
• fix(intelligence): integrate retention fields into runtime paths by @knqiufan in #1084
• fix(intelligence): keep fresh zero-importance memories searchable by @wayyoungboy in #1113
• chore: bump version to 1.1.6 by @wayyoungboy in #1114

New Contributors

• @yihong0618 made their first contribution in #1106
• @kic635 made their first contribution in #1097

Full Changelog: v1.1.5...v1.1.6

PowerMem 1.1.5

What's Changed

• Improve Claude Code plugin init flow by @wayyoungboy in #1024
• fix: clean up graph search tools and results by @wayyoungboy in #1012
• [uv] Use uvx for Claude plugin backend init by @wayyoungboy in #1030
• fix: make graph add relationship writes transactional by @wayyoungboy in #1011
• Fix unit test regressions by @wayyoungboy in #1033
• fix: stabilize regression server restarts by @wayyoungboy in #1034
• build(deps-dev): bump vite from 7.3.2 to 7.3.5 in /dashboard by @dependabot[bot] in #1035
• build(deps): bump launch-editor from 2.12.0 to 2.14.1 in /docs/website by @dependabot[bot] in #1036
• build(deps): bump ws in /docs/website by @dependabot[bot] in #1037
• fix: harden regression server restart handling by @wayyoungboy in #1038
• Support no-LLM mode by @wayyoungboy in #1026
• feat: add NoopEmbedding provider for EMBEDDING_PROVIDER=none by @lightzt99 in #1058
• [codex] Add isolated no-LLM Claude hook regression tests by @wayyoungboy in #1050
• feat: move embedding model download into PowerMem with CN-aware routing by @lightzt99 in #1057
• build(deps): bump webpack-dev-server from 5.2.4 to 5.2.5 in /docs/website by @dependabot[bot] in #1064
• Add Claude Code hook privacy controls by @wayyoungboy in #1054
• fix(graph): add UNIQUE index and upsert for entity dedup (#1003) by @jfeng18 in #1032
• Add coding-agent observation and session timeline APIs by @wayyoungboy in #1051
• Merge MCP dependencies into server extra by @wayyoungboy in #1071
• docs: improve .env.example.full reranker and embedding provider documentation by @knqiufan in #1073
• fix(intelligence): replace unsafe number-grabbing fallback in _parse_importance_response by @knqiufan in #1075
• Publish release plugin and platform binaries by @wayyoungboy in #1060
• fix: resolve docs dependency and checkout deprecation by @wayyoungboy in #1079
• chore: bump version to 1.1.5 by @wayyoungboy in #1083

Full Changelog: v1.1.4...v1.1.5

PowerMem 1.1.4

What's Changed

• fix: handle multimodal image failures and native filter operators by @wayyoungboy in #982
• Fix SDK add when embedded seekdb is owned by server by @wayyoungboy in #973
• Expose filters on AgentMemory retrieval APIs by @wayyoungboy in #976
• ci: tolerate missing regression junit artifacts by @wayyoungboy in #999
• build(deps): bump shell-quote from 1.8.3 to 1.8.4 in /docs/website by @dependabot[bot] in #1000
• doc(claude-code-plugin): improve setup auto-detect and dashboard by @lightzt99 in #994
• Fix Node 20 CI deprecation and VS Code extension cache path by @wayyoungboy in #1005
• fix(intelligence): correct Ebbinghaus decay parameters and lifecycle logic by @jfeng18 in #995
• feat(server): auto-open Dashboard in browser on powermem-server start by @knqiufan in #963
• fix: apply access reinforcement to memory decay by @wayyoungboy in #1006
• fix(skill-store): fix broken embed calls and dedup threshold by @jfeng18 in #1001
• fix(dashboard): show empty states for overview charts by @wayyoungboy in #1007
• docs: add Why PowerMem section by @wayyoungboy in #1009
• fix: make lint usable for critical checks by @wayyoungboy in #1008
• docs: add AgentScope MCP integration guide by @wayyoungboy in #1013
• feat(api): filter listed memories by scope by @wayyoungboy in #1015
• fix(init): include server,seekdb extras and improve LLM auto-detection by @lightzt99 in #1016
• fix: scope graph entities by user by @wayyoungboy in #1010
• Fix search score and threshold handling by @wayyoungboy in #1020
• feat(init): change default data dir to ~/.powermem by @lightzt99 in #1019

New Contributors

• @jfeng18 made their first contribution in #995

Full Changelog: v1.1.3...v1.1.4

v1.1.3

What's Changed

• fix(intelligence): preserve on_get metadata merges on search path by @knqiufan in #981
• Fix agent memory sharing permissions by @wayyoungboy in #972
• feat(search): wire dashboard to semantic search, add time_range + sort by @LINxiansheng in #966
• fix(intelligence): soft mark forgotten memories by @wayyoungboy in #984
• feat(vscode-extension): add CodeFuse IDE support and durable server p… by @happyhust in #986
• Fix PyPI publish trigger for tag releases by @wayyoungboy in #985
• Make soft-forgotten memories rank lower by @wayyoungboy in #987
• chore: add powermem-mcp release flow by @wayyoungboy in #988
• fix(intelligence): expose final search score after ranking by @wayyoungboy in #991
• chore: bump version to 1.1.3 by @wayyoungboy in #990

New Contributors

• @LINxiansheng made their first contribution in #966

Full Changelog: v1.1.2...v1.1.3

v1.1.2

What's Changed

• build(deps-dev): bump vite from 7.3.1 to 7.3.2 in /dashboard by @dependabot[bot] in #896
• build(deps): bump lodash from 4.17.21 to 4.18.1 in /docs/website by @dependabot[bot] in #897
• feat: Add IDE integrations by @Teingi in #365
• docs: rewrite readme by @PsiACE in #374
• Fix SQLite filters key SQL injection (CWE-89) by @WSL0809 in #899
• Action Repair by @Ripcord55 in #900
• Fix #350: [Enhancement]: When some fields in UserProfile are NULL, th... by @JiwaniZakir in #902
• build(deps): bump follow-redirects from 1.15.11 to 1.16.0 in /docs/website by @dependabot[bot] in #903
• Add ob441-reg and sparse case by @Ripcord55 in #905
• feat(examples): Add MoonBit client example for PowerMem HTTP API by @tonyfettes in #906
• fixed(core): OpenAI LLM by @Teingi in #901
• Dashboard Regression Cases by @Ripcord55 in #909
• fix(intelligence): store memory_type in metadata JSON column and enab… by @wayyoungboy in #910
• chore: bump version to 1.1.1 by @wayyoungboy in #912
• feat(config): add env var support for custom extraction/update/import… by @wayyoungboy in #911
• fix(i18n): add missing zh-CN translations for theme switch component by @wayyoungboy in #918
• fix(dashboard): remove incorrect no-data toast description on refresh by @wayyoungboy in #916
• fix(config): resolve sparse_embedder provider from dict on MemoryConf… by @wayyoungboy in #917
• Fixed the issue where a TypeError was thrown when using MemoryGraph due to GraphStoreFactory refusing to accept a complete configuration object. by @wayyoungboy in #919
• build(deps): bump @babel/plugin-transform-modules-systemjs from 7.28.5 to 7.29.4 in /docs/website by @dependabot[bot] in #924
• build(deps): bump qs and body-parser in /docs/website by @dependabot[bot] in #928
• fix(server): load .env before embedded-storage worker detectio by @Teingi in #920
• build(deps): bump postcss from 8.5.6 to 8.5.14 in /docs/website by @dependabot[bot] in #929
• fix(ci): skip scheduled workflows on forked repositories by @wayyoungboy in #944
• feat(deps): split cli/server/mcp/seekdb into optional extras by @wayyoungboy in #931
• build(deps): bump webpack-dev-server from 5.2.2 to 5.2.4 in /docs/website by @dependabot[bot] in #932
• build(deps): bump qs and express in /docs/website by @dependabot[bot] in #934
• fix(intelligence): unify memory-type classification thresholds by @knqiufan in #936
• fix(logging): SDK file logging, trace context, and gzip rotation by @wayyoungboy in #950
• fix(logging): complete logging system gaps across SDK, server, CLI, a… by @wayyoungboy in #952
• fix(user_memory): prevent LLM no-op explanation from overwriting user profile by @lightzt99 in #951
• feat(embedder): zero-config startup with built-in default embedder; split .env.example (closes #940 #941 #949 #948 #) by @happyhust in #945
• docs(integrations): add per-client setup guides and unify MCP default… by @happyhust in #954
• fix(intelligence): unify review-schedule formulas and stored-query behavior by @knqiufan in #938
• build(deps-dev): bump vitest from 3.2.4 to 4.1.0 in /dashboard by @dependabot[bot] in #953
• fix(intelligence): use per-memory decay rate for Ebbinghaus decay by @knqiufan in #956
• fix(deps): make seekdb optional for CentOS 7 compatible installs by @wayyoungboy in #964
• Support custom default headers for OpenAI-compatible providers by @wayyoungboy in #965
• fix(claude-code-plugin): improve SETUP.md + fix Anthropic temperature/top_p conflict by @lightzt99 in #961
• Feature/dashboard documentation 959 by @happyhust in #962
• fix(ci): isolate regression embedding dimensions by @wayyoungboy in #967
• fix(llm): wire base_url into AnthropicLLM and ZaiLLM clients by @lightzt99 in #968
• fix(ci): keep optional dependencies from breaking regression server by @wayyoungboy in #970
• fix(claude-code-plugin): improve SETUP.md — dashboard question, English translations, Python interpreter detection by @lightzt99 in #969
• fix(cli): keep config show section headers for empty optional groups by @wayyoungboy in #974
• Add Claude Code marketplace install flow by @wayyoungboy in #977

New Contributors

• @WSL0809 made their first contribution in #899
• @JiwaniZakir made their first contribution in #902
• @tonyfettes made their first contribution in #906
• @lightzt99 made their first contribution in #951
• @happyhust made their first contribution in #945

Full Changelog: v1.1.0...v1.1.2

v1.1.1

What's Changed

• build(deps-dev): bump vite from 7.3.1 to 7.3.2 in /dashboard by @dependabot[bot] in #896
• build(deps): bump lodash from 4.17.21 to 4.18.1 in /docs/website by @dependabot[bot] in #897
• feat: Add IDE integrations by @Teingi in #365
• docs: rewrite readme by @PsiACE in #374
• Fix SQLite filters key SQL injection (CWE-89) by @WSL0809 in #899
• Action Repair by @Ripcord55 in #900
• Fix #350: [Enhancement]: When some fields in UserProfile are NULL, th... by @JiwaniZakir in #902
• build(deps): bump follow-redirects from 1.15.11 to 1.16.0 in /docs/website by @dependabot[bot] in #903
• Add ob441-reg and sparse case by @Ripcord55 in #905
• feat(examples): Add MoonBit client example for PowerMem HTTP API by @tonyfettes in #906
• fixed(core): OpenAI LLM by @Teingi in #901
• Dashboard Regression Cases by @Ripcord55 in #909
• fix(intelligence): store memory_type in metadata JSON column and enab… by @wayyoungboy in #910

New Contributors

• @WSL0809 made their first contribution in #899
• @JiwaniZakir made their first contribution in #902
• @tonyfettes made their first contribution in #906

Full Changelog: v1.1.0...v1.1.1

v1.1.0

PowerMem 1.1.0

Release date: April 2, 2026

Overview

This minor release focuses on embedded seekdb support for OceanBase-backed deployments, so you can run PowerMem against a local embedded database without a separate database server. It also includes CLI and embedding configuration fixes.

Quick start: from install to embedded seekdb

A minimal runnable path: no Docker, no separate database service—only Python, .env, and your app or CLI.

1. Install

pip install powermem

Embedded seekdb depends on pyseekdb, which is pulled in with powermem; you do not need an extra pip install for it.

2. Configure .env (embedded mode)

Copy the template and edit as needed:

cp .env.example .env

For OceanBase storage with embedded seekdb, the essentials are: do not configure a remote host, and point at a local data directory:

# OceanBase vector storage
DATABASE_PROVIDER=oceanbase

# Embedded mode: leave remote host empty (or unset)
OCEANBASE_HOST=
# Local data directory — same idea as ob_path in config
OCEANBASE_PATH=./seekdb_data

OCEANBASE_PORT=2881
OCEANBASE_USER=root@sys
OCEANBASE_PASSWORD=your_password
OCEANBASE_DATABASE=powermem
OCEANBASE_COLLECTION=memories

# Vector dimension must match your embedding model (example is a common size; adjust for your embedder)
OCEANBASE_EMBEDDING_MODEL_DIMS=1536

Configure LLM and embedding per the README (for example LLM_PROVIDER, LLM_API_KEY, EMBEDDER_*). The embedder output dimension and OCEANBASE_EMBEDDING_MODEL_DIMS must match. You can also run pmem config init for an interactive wizard that explains embedded seekdb vs remote OceanBase.

Added

• Embedded seekdb (OceanBase)
  • Configure a local data directory via ob_path and use embedded mode when no remote host is required.
  • Automatic handling to ensure the target database exists before connecting.
  • Safer defaults for small embedded datasets (e.g. HNSW index behavior where IVF-family indexes are unsuitable).
  • Synchronous update/delete paths for embedded storage to avoid stability issues from concurrent access.
  • pyseekdb dependency for embedded use.

Changed

• CLI (pmem config)

  • Prompts and help text updated to explain embedded seekdb options alongside remote OceanBase.

• OceanBase vector store

  • update merges payloads with existing rows more reliably and reduces accidental loss of fields (including sparse embedding-related data).

Fixed

• CLI

  • --env-file: respects custom env files by loading POWERMEM_ENV_FILE during dotenv initialization.
  • Memory list: consistent ID truncation in non-interactive output; ellipsis for truncated fields in interactive list.

• Embeddings

  • Ollama: ollama_base_url is accepted in Ollama embedding configuration.

Build & tooling

• CI: workflow to export Docker image packages.

Upgrade notes

1. If you use OceanBase, review .env.example for embedded vs remote connection settings (ob_path, host, etc.).
2. For embedded seekdb, use the synchronous Memory API; see the async memory docs for the AsyncMemory limitation.
3. Python 3.11+ remains the supported baseline (as documented in the README).

Full Changelog: v1.0.2...v1.1.0

v1.0.2

What's Changed

• Powermem CLI Cases by @Ripcord55 in #369
• fix(cli): honor --env-file by loading POWERMEM_ENV_FILE in dotenv step by @Teingi in #371

Full Changelog: v1.0.1...v1.0.2