feat: concurrent add pipeline with crash-safe mutation recovery

README.md

@@ -333,6 +333,7 @@ OpenKB settings are initialized by `openkb init` and stored in `.openkb/config.y
 model: gpt-5.4                   # LLM model (any LiteLLM-supported provider)
 language: en                     # Wiki output language
 pageindex_threshold: 20          # PDF pages threshold for PageIndex
+file_processing_jobs: 2          # Files to prepare concurrently during `openkb add <dir>`
 ```
 
 Model names use `provider/model` LiteLLM [format](https://docs.litellm.ai/docs/providers) (OpenAI models can omit the prefix):
@@ -347,6 +348,8 @@ Model names use `provider/model` LiteLLM [format](https://docs.litellm.ai/docs/p
 <summary><i>Advanced options (entity_types, extra_headers, OAuth):</i></summary>
 <br>
 
+`file_processing_jobs` (default `2`): number of files prepared concurrently during `openkb add <dir>`. Only the preparation stage is parallelized (hashing, duplicate prefiltering, raw/source staging, conversion); live-KB mutation stays serialized under the mutation lock, so raising it helps mainly when conversion is the bottleneck.
+
 `entity_types` (optional): a YAML list overriding the entity-type vocabulary used for entity pages; omit it to use the default `person`, `organization`, `place`, `product`, `work`, `event`, `other`.
 
 `extra_headers` (optional): a YAML mapping of extra HTTP headers sent with every LLM request (forwarded to LiteLLM's `extra_headers`). Useful for providers that expect custom headers, e.g. GitHub Copilot IDE-auth headers:

config.yaml.example

@@ -1,6 +1,10 @@
 model: gpt-5.4                   # LLM model (any LiteLLM-supported provider)
 language: en                     # Wiki output language
 pageindex_threshold: 20          # PDF pages threshold for PageIndex
+file_processing_jobs: 2          # Number of files to prepare concurrently during `openkb add <dir>`
+# Note: this parallelizes hashing/conversion/staging only. Live KB publish,
+# PageIndex indexing, LLM compilation, registry updates, and log writes remain
+# serialized under the KB mutation lock.
 
 # Optional: extra HTTP headers sent with every LLM request (forwarded to
 # LiteLLM's extra_headers). Some providers need these — e.g. GitHub Copilot